Cell quality control mechanisms maintain stemness and differentiation potential of P19 embryonic carcinoma cells

ABSTRACT

Given the relatively long life of stem cells (SCs), efficient mechanisms of quality control to balance cell survival and resistance to external and internal stress are required. Our objective was to test the relevance of cell quality control mechanisms for SCs maintenance, differentiation and resistance to cell death. We compared cell quality control in P19 stem cells (P19SCs) before and after differentiation (P19dCs). Differentiation of P19SCs resulted in alterations in parameters involved in cell survival and protein homeostasis, including the redox system, cardiolipin and lipid profiles, unfolded protein response, ubiquitin-proteasome and lysosomal systems, and signaling pathways controlling cell growth. In addition, P19SCs pluripotency was correlated with stronger antioxidant protection, modulation of apoptosis, and activation of macroautophagy, which all contributed to preserve SCs quality by increasing the threshold for cell death activation. Furthermore, our findings identify critical roles for the PI3K-AKT-MTOR pathway, as well as autophagic flux and apoptosis regulation in the maintenance of P19SCs pluripotency and differentiation potential.

Abbreviations: 3-MA: 3-methyladenine; AKT/protein kinase B: thymoma viral proto-oncogene; AKT1: thymoma viral proto-oncogene 1; ATG: AuTophaGy-related; ATF6: activating transcription factor 6; BAX: BCL2-associated X protein; BBC3/PUMA: BCL2 binding component 3; BCL2: B cell leukemia/lymphoma 2; BNIP3L: BCL2/adenovirus E1B interacting protein 3-like; CASP3: caspase 3; CASP8: caspase 8; CASP9: caspase 9; CL: cardiolipin; CTSB: cathepsin B; CTSD: cathepsin D; DDIT3/CHOP: DNA-damage inducible transcript 3; DNM1L/DRP1: dynamin 1-like; DRAM1: DNA-damage regulated autophagy modulator 1; EIF2AK3/PERK: eukaryotic translation initiation factor 2 alpha kinase 3; EIF2S1/eIF2α: eukaryotic translation initiation factor 2, subunit alpha; ERN1/IRE1α: endoplasmic reticulum to nucleus signaling 1; ESCs: embryonic stem cells; KRT8/TROMA-1: cytokeratin 8; LAMP2A: lysosomal-associated membrane protein 2A; MAP1LC3/LC3: microtubule-associated protein 1 light chain 3; MTOR: mechanistic target of rapamycin kinase; NANOG: Nanog homeobox; NAO: 10-N-nonyl acridine orange; NFE2L2/NRF2: nuclear factor, erythroid derived 2, like 2; OPA1: OPA1, mitochondrial dynamin like GTPase; P19dCs: P19 differentiated cells; P19SCs: P19 stem cells; POU5F1/OCT4: POU domain, class 5, transcription factor 1; PtdIns3K: phosphatidylinositol 3-kinase; RA: retinoic acid; ROS: reactive oxygen species; RPS6KB1/p70S6K: ribosomal protein S6 kinase, polypeptide 1; SCs: stem cells; SOD: superoxide dismutase; SHC1-1/p66SHC: src homology 2 domain-containing transforming protein C1, 66 kDa isoform; SOX2: SRY (sex determining region Y)-box 2; SQSTM1/p62: sequestosome 1; SPTAN1/αII-spectrin: spectrin alpha, non-erythrocytic 1; TOMM20: translocase of outer mitochondrial membrane 20; TRP53/p53: transformation related protein 53; TUBB3/betaIII-tubulin: tubulin, beta 3 class III; UPR: unfolded protein response; UPS: ubiquitin-proteasome system

Introduction

Stem cells (SCs) balance resistance and susceptibility to cell death by multiple pathways including a high mitochondrial priming and proximity to the apoptotic threshold, intimately linked to pluripotency and differentiation [1]. Although tissue engineering, regenerative medicine and stem cell-based therapies are frequently focused on differentiation protocols, the mechanisms that regulate cell death and survival in SCs are not completely understood, and are particularly relevant for the development of more effective interventional strategies. Recent work has demonstrated that metabolic status critically impacts not only the survival of SCs, but also their differentiation potential and fate [2].

Embryonal carcinoma cells, such as the P19 cell line, are considered the malignant counterparts of embryonic stem cells (ESCs) and constitute an excellent model to study SCs maintenance and differentiation. Similarly to ESCs, P19 stem cells (P19SCs) are able to differentiate in the presence of retinoic acid (RA) depending on dosage and growth conditions [3]. At low concentrations, RA directs P19 cells to differentiate into an endodermal phenotype, whereas at higher concentrations RA induces P19 differentiation to neuroectoderm. Due to this, P19 cells are one of the most studied in vitro systems to investigate RA-induced differentiation that is mainly mediated by specific nuclear receptors (RAR) [4]. Thus, P19 cells grown in monolayer and treated with 1 μM RA yields a mixed population of P19 differentiated cells (P19dCs) with endodermal and neuroectodermal phenotypes after 4 days [5,6]. This differentiation pattern is also characterized by the appearance of more lobular and euchromatic nuclei, cell flattening and alterations in microfilament organization. Interestingly, we previously found that the differentiation of P19 cells was accompanied by mitochondrial remodeling [6], showing a unique association between mitochondrial activity, cell differentiation, and stemness. In comparison with their differentiated counterparts, P19SCs pluripotency was correlated with a strong glycolytic profile and decreased mitochondrial biogenesis and complexity: round, weakly-polarized and inactive mitochondria with a closed permeability transition pore. This lower mitochondrial activity increased P19SCs resistance against dichloroacetate. Thus, stimulation of mitochondrial function by growing P19SCs in glutamine/pyruvate (glucose-free)-containing medium reduced their glycolytic phenotype, induced loss of pluripotency, compromised differentiation, and increased the susceptibility of P19SCs to dichloroacetate [6]. In addition, we found that the more glycolytic and undifferentiated cells were less susceptible to melatonin [7]. This pineal hormone wields its antiproliferative effects only in differentiated cells with an active oxidative metabolism, triggering a form of mitochondria-mediated cell death which is characterized by an arrest at S-phase, reduction of the mitochondrial electron transport chain, generation of reactive oxygen species (ROS), BCL2 (B-cell leukemia/lymphoma 2) downregulation and apoptosis-inducing factor release [7]. These findings highlight the importance of mitochondrial metabolism in stemness, proliferation, differentiation, and resistance to cell death [8]. However, little is known about the role of redox system and cell quality control mechanisms, including apoptosis and autophagy, which coordinate cell growth and metabolism with death and survival decisions, in SCs maintenance and differentiation.

Several studies [9–12] have suggested a close relationship between mitochondrial function, phospholipid composition and peroxidation, and cell life and death decisions. In addition, cell quality control systems facilitate an efficient degradation of damaged proteins, structures or cells and play a key role in retarding cellular senescence and tissue dysfunction [13–15]. The unfolded protein response (UPR), together with autophagy, is considered a major player in cell quality control and maintenance of cellular homeostasis, as well as in cellular remodeling during normal development [16]. Furthermore, several signal-transduction cascades interact with the autophagy machinery in order to detect fluctuations in key metabolic and redox parameters and adapt cells to adverse conditions, including limited nutrient supply or oxygen deprivation [17]. Given the relatively long life of SCs in different tissues [18], efficient mechanisms of quality control to balance cell survival and resistance to damage are important, thus preserving the SCs pool in its particular microenvironmental niche. Although recent research suggests that autophagy impacts SCs development [19], studies focusing on the role of cell quality control systems in self-renewal, stemness maintenance, cell differentiation, and resistance to cell death are scarce. Our hypothesis is that SCs maintenance and survival is only guaranteed by assuring prompt corrective actions to prevent, suppress or tolerate any stress or damage. In support of this, we find better cell quality parameters in P19SCs than P19dCs, and that a close relationship exists between pluripotency and cell quality control systems, thus impacting SCs fate.

Results

P19SCs present strong antioxidant protection

Differentiation of P19SCs with RA increased both intracellular reactive oxygen and nitrogen species. We also identified an alteration of the enzymatic antioxidant defense after the differentiation protocol, with higher SOD (superoxide dismutase) and lower catalase activities, an effect probably driven by the high levels of mitochondrial superoxide anion in P19dCs [6]. Despite this, P19SCs showed a strong antioxidant protection characterized by elevated concentrations of GSH and higher total antioxidant capacity (Figure 1(a)). This particular redox status found in P19SCs is intimately linked with their lower mitochondrial activity. In fact, P19SCs grown in galactose (glucose-free)-containing media to force mitochondrial energy metabolism led to a reduction in GSH:GSSG ratio (Fig. S1). In addition, although protein carbonyl levels remained unaltered, P19dCs showed more extensive lipid peroxidation (Figure 1(a)).

Figure 1.

P19SCs differentiation involves increased intracellular oxidative stress. (a) Confocal images obtained using H₂DCFDA as an indicator for reactive oxygen species, scale bar: 20 μm. Levels of reactive nitrogen species (RNS), catalase (CAT) and superoxide dismutase (SOD) activities, reduced (GSH) and oxidized (GSSG) glutathione, total antioxidant activity, carbonylated proteins and lipid peroxidation were also analyzed. Data are means ± S.D., n = 5. Significant p-values from Student’s t-test are indicated in the graph. (b) Protein levels of mitochondrial superoxide dismutase (SOD2), NFE2L2, SHC1 isoforms SHC1-1/p66SHC, SHC1-2/p52SHC and SHC1-3/p46SHC and p-SHC1-1 (S36) in P19SCs and P19dCs. Bar charts show means of optical density (O.D.) ± S.D. expressed as percentage of P19SCs, from three separate immunoblots. Ponceau S was used for gel loading control. Significant p-values from Student’s t-test are indicated in the graph.

Other parameter linking the low mitochondrial activity of P19SCs with their antioxidant protection was the increased content of SOD2. Thus, P19SCs are probably prepared to trigger more efficient antioxidant responses against oxidative insults. Accordingly, we found that undifferentiated cells showed higher NFE2L2/NRF2 (nuclear factor, erythroid derived 2, like 2) expression, a critical factor for promoting antioxidant responses and cell survival. In fact, P19SCs maintained higher proliferation rates when treated with H₂O₂ than P19dCs and P19 cells grown in galactose (glucose-free)-containing media which rely more in mitochondrial metabolism for ATP production (Fig. S2). At the same time, although differentiated cells showed higher oxidative stress and lipid peroxidation, we found a higher expression of SHC1-1/p66SHC (src homology 2 domain-containing transforming protein C1, 66 kDa isoform) and its phosphorylated form at Ser36 in P19SCs, suggesting an activation of this pro-oxidant and pro-apoptotic signaling pathway in undifferentiated cells (Figure 1(b)).

Differentiation of P19SCs alters the lipid profile and remodels cardiolipin

Since lipids can act as alternative energy sources and signaling entities, lipid metabolism has been shown to influence cell proliferation, differentiation and death, thus representing new possible operators in SCs regulation [20]. Despite this, we found that the phospholipid profile was relatively similar in both types of P19 cells. In both P19SCs and P19dCs, phosphatidylcholine represented the dominant phospholipid class of total phospholipids followed by phosphatidylethanolamine. Additionally, other phospholipids were also detected with the following order of abundance: phosphatidylserine, phosphatidylinositol and cardiolipin (CL) in P19SCs and phosphatidylinositol, phosphatidylserine and CL in P19dCs (Figure 2(a)). No significant variations between groups were observed except for phosphatidylethanolamine whose content was increased after the treatment with RA. Therefore, phosphatidylcholine:phosphatidylethanolamine ratio, which was shown to play a role in maintaining membrane integrity [21], showed lower values in P19dCs (3.591 ± 0.395 in P19SCs and 2.144 ± 0.622 in P19dCs, p < 0.01).

Figure 2.

P19SCs differentiation remodels phospholipid profile. (a) Phospholipid content of P19SCs and P19dCs. PC – phosphatidylcholine, PI – phosphatidylinositol, PS – phosphatidylserine, PE – phosphatidylethanolamine, CL- cardiolipin. Relative abundance of phospholipid refers to the percentage of phospholipid phosphorus recovered from respective thin-layer chromatography. Data represent the average percentage of phospholipid content ± SEM from 5 independent experiments. Significant p-values from Student’s t-test are indicated in the graph. (b) Cardiolipin peroxidation was monitored by following the loss of NAO fluorescence by flow cytometry using H₂O₂ as positive control. Demonstrative histograms show the loss of NAO fluorescence peak towards the increase of M1 region peak. The % cells gated in M1 region indicates increased of cardiolipin oxidation. Data are means ± S.D. from n = 6. Significant p-values from two-way analysis of variance and post hoc comparisons are displayed in the graph.

Due to the role of CL in mitochondrial function and cell death [22], we analyzed the CL molecular species by mass spectrometry to confirm the absence of significant changes among groups. CL has four alkyl groups and two phosphate groups that makes CL able to potentially possess two negative charges resulting in either singly charged [M-H]⁻ or doubly charged [M-2H]²⁻ ions observable in the mass spectrometry spectra [23]. Singly charged CL ions in negative mode were represented by different molecular clusters with m/z 1373.8, 1401.7, 1403.8, 1425.8, 1427.8, 1449.8, 1455.8, 1451.8, 1453.8, 1479.8, 1501.8 with a variety of fatty acid residues (from C16:0 to C22:6). Figure S3A shows the mass spectra of [M-H]⁻ ions of CL obtained from both types of P19 cells. The molecular ion clusters included CL molecular species at m/z 1401.8; 1427.8; 1455.8; 1449.8; 1479.8. The ions observed in the spectra and attributed to CL were identified as described in Table 1.

Table 1.

Major cardiolipin (CL) molecular species from P19SCs and P19dCs.

CL		Diacyl species
[M-H]−	[M-2H]2−	Fatty acyl composition
1401.7	701.5	(16:0)2(18:1)(18:2); (16:0)(16:1)(18:1)2
1403.8	702.5	(16:0)2(18:1)2
1425.8	712.5	(16:0)(18:1)(18:2)2
1427.8	713.5	(16:0)(18:1)2(18:2)
1449.8	724.5	(18:1)(18:2)3; (16:0)(18:2)2(20:4)
1451.8	725.5	(18:1)2(18:2)2
1453.8	726.5	(18:1)3(18:2)
1455.8	727.5	(18:1)4; (16:1)(18:0)(18:1)(20:2)
1479.8	739.5	(18:1)3(20:3); (18:0)(18:1)2(20:4)
1501.8	750.5	(20:4)(20:3)(18:1)2; (22:6)(18:1)3

The results show evident differences in CL species between both types of P19 cells. The CL species at m/z 1427.8 and 1455.8, attributed to the CL (16:1)(18:2)(18:1)₂ and CL (18:1)4 and (16:1)(18:0)(18:1)(20:2) respectively, were highly present in P19SCs; while in P19dCs, CL species at m/z 1427.5 corresponding to CL (16:0)(18:2)(18:1)₂ was the most abundant species. The relative abundance of the ion at m/z 1403.8 attributed to the CL (16:0)₂(18:1)₂ had similar intensities in both types of P19 cells. Overall, P19SCs were enriched with CL species at m/z 1425.8 and 1449.8 corresponding to CL (16:0)(18:1)(18:2)₂ and (18:1)(18:2)₃ and (16:0)(18:2)2(20:4) respectively, which were absent in P19dCs (Fig. S3A). This remodeling in CL was confirmed by a higher expression of CRLS1 (cardiolipin synthase 1) and TAZ/tafazzin in more differentiated cells (Fig. S3B). In addition, we also analyzed CL oxidation using the fluorescent probe 10-N-nonyl acridine orange (NAO) and H₂O₂ as positive control. As shown in Figure 2(b), the loss of NAO fluorescence, what reflects higher CL peroxidation, is shown as a disappearance of the NAO peak signal from region M2 to region M1. Our results shown as the percentage of cells with oxidized CL (cells gated on M1 region) revealed evident differences between undifferentiated and differentiated cells. Thus, P19SCs showed lower levels of peroxidized CL. The treatment with H₂O₂ increased the number of cells in the M1 region in both groups of P19 cells. However, the effect of H₂O₂ on CL peroxidation was more limited in P19dCs probably because their CL has a higher basal degree of oxidation.

P19SCs present alterations in the apoptotic machinery

We next studied some components of the apoptotic pathway to detect the alterations produced in its machinery after the differentiation of P19SCs that may further explain their particular resistance to cell death [6,7]. We found an overexpression of the tumor suppressor TRP53/p53 (transformation related protein 53) and its transcriptional targets BBC3/PUMA (BCL2 binding component 3) and DRAM1 (Damage-regulated autophagy modulator 1) in P19SCs, suggesting the activation of cell death processes that decrease after differentiation with RA. Despite this, BAX (BCL2-associated X), another TRP53-inducible protein, and BCL2 protein levels showed no statistical differences between both types of P19 cells, while BNIP3L (BCL2/adenovirus E1B 19 kDa interacting protein 3-like), a BCL2-related protein with cell death and autophagic functions [24], displayed a higher protein expression in P19dCs. However, although P19SCs presented a higher content in cleaved CASP3 (caspase 3) than P19dCs, this was not associated with cleavage of classical substrates such as SPTAN1/αII-spectrin (spectrin alpha, non-erythrocytic 1). We only detected the band corresponding to the calpain-cleaved product of SPTAN1 in P19dCs (Figure 3(a)). Indeed, CASP8 (caspase 8)-, CASP9 (caspase 9)- and CASP3-like activities (Figure 3(b)) and the percentage of dead cells (calcein⁻ propidium iodide⁺) (Figure 3(c)) were increased after the differentiation with RA, indicating a suppression of apoptotic signaling in undifferentiated cells that probably constitutes a strategy to increase the thresholds for triggering cell death programs.

Figure 3.

P19SCs-P19dCs transition alters the apoptotic machinery. (a) Semi-quantification of protein content by immunoblotting of TRP53, BBC3, DRAM1, BAX, BCL2, BNIP3L, cleaved CASP3 and SPTAN1 cleavage products in P19SCs and P19dCs. Bar charts show means of optical density (O.D.) ± S.D. expressed as percentage of P19SCs, from three separate experiments. Ponceau S was used for loading control. Significant p-values from Student’s t-test are indicated in the graph. (b) CASP8-, CASP9- and CASP3-like activities determined in P19SCs and P19dCs using specific substrates. Data are expressed as means ± S.D. from 5 independent experiments. Significant p-values from Student’s t-test are indicated in the graph. (c) Live/dead assay with calcein-AM and propidium iodide (PI) in P19SCs and P19dCs. Data are expressed as percentage of viable cells (calcein⁺ PI⁻), dead cells (calcein⁻ PI⁺), and cell debris (calcein⁻ PI⁻) ± SD from at least 3 independent experiments. Significant p-values from Student’s t-test are indicated in the graph.

P19SCs differentiation activates ATF6 and EIF2S1 arms of the UPR and unfolded monomer proteins degradation systems

The UPR adapts cells against the stress of misfolded proteins in the endoplasmic reticulum and, when this adaptation fails, it is also able to trigger apoptosis. In fact, glucose deprived cancer cells undergo cell death by UPR-mediated mechanisms [25]. Since the UPR may play a role during metabolic reprogramming accompanying SCs differentiation, we analyzed the expression of three unfolded protein sensors in both types of P19 cells. ERN1/IRE1α (endoplasmic reticulum to nucleus signaling 1) senses ER unfolded proteins and showed a lower protein level after RA-induced differentiation. Despite this, EIF2AK3/PERK (eukaryotic translation initiation factor 2 alpha kinase 3)-EIF2S1 (eukaryotic translation initiation factor 2, subunit alpha) and ATF6 (activating transcription factor 6) arms of the UPR that coordinates rRNA transcription and translation inhibition, promotes protein degradation, and usually predisposes to apoptosis or autophagy, were activated in P19dCs. Despite this, DDIT3/CHOP (DNA-damage inducible transcript 3), a key mediator of the ER stress-mediated apoptosis pathway, showed a decreased protein content in P19dCs (Figure 4(a)).

Figure 4.

Differentiation of P19SCs activates unfolded protein responses (UPR), proteasome-ubiquitin system and chaperone-mediated autophagy (CMA). (a) Levels of key proteins involved in the three arms of the UPR: ERN1, p-EIF2S1 (S51), ATF6 and DDIT3 in P19SCs and P19dCs. Bar charts show means of optical density (O.D.) ± S.D. expressed as percentage of P19SCs, from three separate experiments. Ponceau S was used for gel loading control. Significant p-values from Student’s t-test are indicated in the graph. (b) Ubiquitinated proteins and the marker for CMA function, LAMP2A, were semi-quantified by immunoblotting in both types of P19 cells. Bar charts show means of optical density (O.D.) ± S.D. expressed as percentage of P19SCs, from three separate experiments. Ponceau S was used for loading control. Significant p-values from Student’s t-test are indicated in the graph. (c) Chymotrypsin-like activity of the 20S proteasome was measured in undifferentiated and differentiated P19 cells. Lactacystin was used to specifically block proteasomal degradation. Data are means ± S.D. from n = 6. Significant p-values from two-way analysis of variance and post hoc comparisons are displayed in the graph.

ER stress can be ameliorated by the degradation of misfolded and unfolded proteins. Accordingly with the preceding results suggesting the accumulation of unfolded/misfolded proteins in P19dCd, the differentiated cells also presented higher LAMP2A (lysosomal-associated membrane protein 2A) expression, and increased protein ubiquitination (Figure 4(b)) and 20S proteasome activity (Figure 4(c)). Therefore, chaperone-mediated autophagy and the ubiquitin-proteasome system (UPS) appear to be upregulated during the process of P19SCs differentiation.

P19SCs differentiation induces a remodeling of lysosomal compartments and activates AKT/MTOR pathway

To further explore proteases activation during P19SCs differentiation, the lysosomal proteases CTSD (cathepsin D) and CTSB (cathepsin B) were also measured. CTSD processing seemed to be favored after differentiation, increasing its activity. Furthermore, CTSB activity and the CTSB:CTSD ratio were also higher in differentiated than in undifferentiated cells (Figure 5(a)). LysoTracker dye staining (Figure 5(b)) and electron microscopy (Figure 5(c)) demonstrated a more developed endosomal-lysosomal compartment in P19dCs. Thus, multiple systems for protein degradation seem to be activated after P19SCs differentiation. At the same time, however, AKT1 (thymoma viral proto-oncogene 1) phosphorylated at Ser473, MTOR (mechanistic target of rapamycin kinase) phosphorylated at Ser2448 and RPS6KB1/p70S6K (ribosomal protein S6 kinase, polypeptide 1) quantity was increased in P19dCs (Figure 5(d)), indicating activation of this survival pathway, and suggestive of an increase in protein synthesis during differentiation. These data point to a strong protein turnover and remodeling during RA-induced differentiation of P19 cells.

Figure 5.

Differentiation of P19SCs adapts the endosomal-lysosomal system and activates AKT/MTOR pathway. (a) Immunoblot images to follow the proteolytic maturation of CTSD in P19SCs and P19dCs. Bar charts show means of optical density from bans corresponding to 45 kDa pro-CTSD (O.D.) ± S.D. expressed as percentage of P19SCs, from three separate experiments. Ponceau S was used as gel loading control. Lysosomal proteases CTSD and CTSB activities and CTSB:CTSD ratio were also analyzed in both types of P19 cells. Data are means ± S.D. from n = 6. Significant p-values from Student’s t-test are indicated in the graph. (b) Confocal images obtained using LysoTracker for labeling lysosomal compartments show changes in the presence of acidic organelles between both types of P19 cells. Scale bar = 20 μm. (c) Electron micrographs of P19SCs and 19dCs show a higher content in single-membrane vesicles. Note the presence a vesicle containing an organelle-like structure in P19dCs. Pictures show representative fields of over 50 cells photographed. Scale bar = 3 μm. (d) Protein levels of p-AKT1 (S473), AKT, p-MTOR (S2448), MTOR and RPS6KB1 in P19SCs and P19dCs. Bar charts show means of optical density (O.D.) ± S.D. expressed as percentage of P19SCs, from three separate immunoblots. Ponceau S was used as gel loading control. Significant p-values from Student’s t-test are indicated in the graph.

P19SCs present higher autophagic flux than P19dCs

To evaluate whether cellular quality control by macroautophagy is upregulated during P19SCs differentiation, we measured the expression of key AuTophaGy-related (ATG) proteins in both types of P19 cells. BECN1/Beclin-1, a core component of the class III phosphatidylinositol 3-kinase (PtdIns3K) complex that initiates phagophore formation, showed a higher expression in P19SCs, while ATG12 and ATG12–ATG5 complex, a ubiquitin-protein ligase (E3)-like enzyme for MAP1LC3/LC3 (microtubule-associated protein 1 light chain 3)–phosphatidylethanolamine conjugation reaction, showed higher expression in P19dCs. In addition, ATG3, MAP1LC3-II, a marker of autophagosomes, and MAP1LC3-II:MAP1LC3-I ratio remained unaltered during differentiation. To dissect these apparently contradictory results, we measured the global autophagic flux. Immunoblot analysis of SQSTM1/p62 (sequestosome 1) revealed the accumulation of this protein in P19dCs (Figure 6(a)). Increased SQSTM1 may be due to a transcriptional response usually mediated by NFE2L2 [26] in response to increased ROS or to a block in autophagy. Thus, to evaluate autophagy flux, we inhibited autophagosome-lysosome fusion with bafilomycin A1 and detected MAP1LC3-II accumulation that was higher in P19SCs (Figure 6(b)), confirming a lower autophagic flux in P19dCs. Therefore, P19SCs presented higher rates of autophagy than their differentiated counterparts, which may be related to organelle quality control over periods of relative metabolic quiescence. These changes in the autophagic flux between both types of P19 cells also resulted in divergent effects on cell proliferation when cells were treated with the phosphatidylinositol 3-kinase inhibitor 3-methyladenine (3-MA). P19SCs drastically reduced their proliferation rate while P19dCs maintained their viability. On the contrary, when both types of cells were treated with the MTOR inhibitor rapamycin or transfected with Atg7 siRNA oligonucleotide (si-Atg7) to inhibit autophagy machinery, cell proliferation was decreased only in P19dCs (Figure 6(c)). This indicates that PtdIns3K signaling is necessary for P19SCs maintenance while MTOR activation and intact autophagy machinery are essential for maintaining the proliferation rate of differentiated cells.

Figure 6.

Undifferentiated P19 cells present higher autophagic flux than RA-differentiated cells. (a) Immunoblot analysis of the autophagic markers BECN1, ATG3, ATG12–ATG5 complex, ATG12, MAP1LC3-I, MAP1LC3-II and SQSTM1/p62 in both types of P19 cells. Bar charts show means of optical density (O.D.) ± S.D. expressed as percentage of P19SCs, from three separate immunoblots. Ponceau S was used for loading control. Significant p-values from Student’s t-test are displayed in the graph. (b) Expression of the autophagosomal marker MAP1LC3-II in P19SCs and P19dCs after blocking late-phase autophagy with bafilomycin A1. Bar charts show means of optical density (O.D.) ± S.D. expressed as percentage of P19SCs, from three separate immunoblots. Ponceau S was used for gel loading control. Significant p-values from two-way analysis of variance and post hoc comparisons are indicated in the graph. (c) Effect of the phosphatidylinositol 3-kinase inhibitor 3-methyladenine (3-MA) at 2.5 and 5 mM, the MTOR inhibitor rapamycin at 0.1, 10 and 100 nM, and Atg7 silencing (si-Atg7) on cell proliferation of both types of P19 cells after 72 hours treatment. Data are expressed as percentage of the control (vehicle only or scrambled siRNA (si-Control)). Data are means ± S.E.M. from n = 5. Significant p-values from analysis of variance and post hoc comparisons are displayed in the graph.

Imbalanced mitochondrial dynamics has been associated with altered mitochondrial degradation, resulting in the appearance of elongated and dysfunctional mitochondria. P19 differentiation induced a higher content of fusion-related protein OPA1 (OPA1, mitochondrial dynamin like GTPase) in whole-cell extracts while fission-related protein DNM1L/DRP1 (dynamin 1-like) remained unaltered (Figure 7(a)), suggesting changes in mitochondrial turnover rates. In fact, after treatment with bafilomycin A1 to measure mitophagic flux, the accumulation of SQSTM1-bound mitochondria was higher in P19SCs (Figure 7(b)) indicating a higher mitochondrial turnover in undifferentiated cells. However, additional assays using tandem-tagged mitophagy reporters are still required to further confirm higher mitophagy levels in P19SCs. Autophagy in P19SCs presumably contributes to the maintenance and selection of mitochondrial subpopulations with a proper quality and metabolism for maintaining pluripotency.

Figure 7.

Differentiation of P19SCs promotes mitochondrial fusion and priming for mitophagy. (a) Levels of mitochondrial fission protein DNM1L and fusion protein OPA1 in P19SCs and P19dCs. Bar charts show means of optical density (O.D.) ± S.D. expressed as percentage of P19SCs, from three separate immunoblots. Ponceau S was used as gel loading control. Significant p-value from Student’s t-test is indicated in the graph. (b) Protein expression of SQSTM1/p62 in mitochondrial extracts after blocking late-phase autophagy with bafilomycin A1 shows higher mitophagic flux in P19SCs. TOMM20 was used as loading control. Bar chart displays means of optical density (O.D.) ± S.D. (normalized to TOMM20) expressed as percentage of mitochondrial SQSTM1/p62 detected in untreated P19SCs (P19SCs Control), from three separate experiments. Significant p-values from two-way analysis of variance and post hoc comparisons are displayed in the graph.

PtdIns3K signaling, autophagy and apoptosis are required for stemness maintenance and differentiation in P19 cells

Inhibition of PtdIns3K, autophagy and apoptosis impacted pluripotency of P19SCs. Immunoblot experiments revealed a loss in POU5F1/OCT4 (POU domain, class 5, transcription factor 1) and SOX2 (SRY (sex determining region Y)-box 2) in P19SCs treated with the PtdIns3K inhibitor wortmannin (Figure 8(a)) or with siRNA targeting Atg7 (Figure 8(b)) while the caspases inhibitor Z-VAD-FMK reduced POU5F1 expression without altering NANOG (Nanog homeobox) and SOX2 expression (Figure 9). In addition, wortmannin and Z-VAD-FMK drastically increased TUBB3/betaIII-tubulin (tubulin, beta 3 class III) expression (Figures 8(a) and 9) suggesting that PtdIns3K and apoptosis inhibition trigger SCs differentiation.

Figure 8.

PtdIns3K and autophagy inhibition alters cell fate. (a) Protein levels of pluripotency (POU5F1, NANOG and SOX2) and differentiation (KRT8 and TUBB3) markers in P19 stem cells (P19SCs) treated with the PtdIns3K inhibitor wortmannin (100 nM) and P19 cells differentiated with retinoic acid (P19dCs) in the presence of 100 nM wortmannin. Bar charts show means of optical density (O.D.) ± S.D. expressed as percentage of untreated P19SCs, from three separate immunoblots. Ponceau S was used for gel loading control. Significant p-values from two-way analysis of variance and post hoc comparisons are displayed in the graph. (b) Protein levels of pluripotency (POU5F1, NANOG and SOX2) and differentiation (KRT8 and TUBB3) markers in P19SCs transfected with either Atg7 siRNA oligonucleotide (si-Atg7) or with a scrambled siRNA (si-Con) and in P19 cells differentiated in the presence of siRNA oligonucleotides. Bar charts show means of optical density (O.D.) ± S.D. expressed as percentage of untransfected P19SCs (P19SCs Con), from three separate immunoblots. ATG7 protein levels were detected to confirm the persistence of protein silencing and Ponceau S was used as gel loading control. Significant p-values from two-way analysis of variance and post hoc comparisons are displayed in the graph.

Figure 9.

Caspases inhibition alters cell fate. Protein levels of pluripotency (POU5F1, NANOG and SOX2) and differentiation (KRT8 and TUBB3) markers in P19SCs treated with the pan-caspase inhibitor Z-VAD-FMK (50 μM) and P19 cells differentiated in the presence of 50 μM Z-VAD-FMK. Bar charts show means of optical density (O.D.) ± S.D. expressed as percentage of untreated P19SCs, from three separate immunoblots. Ponceau S was used as gel loading control. Significant p-values from two-way analysis of variance and post hoc comparisons are displayed in the graph.

To assess whether PtdIns3K signaling, autophagy and apoptosis are indispensable for cell differentiation, P19SC were differentiated in the presence of wortmannin, Atg7 siRNA oligonucleotide and Z-VAD-FMK and changes in the expression pattern of differentiation markers were evaluated. Although the treatments with wortmannin and Z-VAD-FMK slightly decreased TUBB3 expression in P19dCs, the trophoectodermal marker KRT8/cytokeratin 8 (TROMA-1 antibody) was significantly reduced (Figures 8(a) and 9). In addition, Atg7 silencing reduced TUBB3 and KRT8 protein expression levels of P19dCs, confirming the disruption in the normal differentiation pattern induced by RA (Figure 8(b)). Thus, PtdIns3K signaling, autophagic and apoptotic pathways seems to be necessary to maintain pluripotency and induce a proper differentiation rate.

Discussion

We investigated the relevance of cell quality control mechanisms, as well as signaling pathways related with cell death and survival, for the differentiation of P19 cells to better understand the interaction between the mechanisms of protein homeostasis such as autophagy and UPR with mitochondrial metabolism, cell survival and pluripotency in SCs.

Our data showed a stronger antioxidant protection on P19SC compared with their differentiated counterparts, and the activation of a pro-oxidant environment with their differentiation. We have previously reported alterations on P19SC differentiation markers when cells were differentiated in the presence of the antioxidant N-acetylcysteine that retained P19 cells in a trophoectodermal stage [6]. Notwithstanding, the status of the redox signaling pathways NFE2L2 and SHC1-1 highlight the delicate equilibrium of SCs to balance resistance and sensitivity to cell death. Despite the amount of research confirming the pro-apoptotic functions of SHC1-1, some studies attribute a role of this pathway in promoting SCs survival under hypoxia, and in maintaining self-renewal [27]. Thus, it was suggested that SHC1-1 could act as a ‘double-edged sword’ in the regulation of cell death, based on microenvironmental conditions, redox status and metabolic context [28], playing a pivotal role in coordinating survival autophagy, autophagic cell death and apoptosis, depending on the bioenergetic status. In fact, aberrant levels of TRP53 and SHC1-1 were detected in cancer stem cells and differentiated counterparts facilitating the acquisition of resistance phenotypes [29,30], a mechanism that probably also operates in a sub-set of pluripotent cells.

Oxidative damage to lipids accompanies differentiation of P19 cells and our results also suggest that CL species and fatty acid composition are altered during SCs differentiation. Thus, we found higher levels of phosphatidylethanolamine in P19dCs. It was described that this elevation is an early but not essential event for P19 cell differentiation into cardiac myocytes [31]. Although no changes in total CL levels were detected, we found evident alterations in enzymes involved in CL synthesis and maturation that were correlated with changes in CL species, and that are closely related with the degree of differentiation and mitochondrial activity. Therefore, this CL remodeling may have important physiological implications by contributing to the mitochondrial remodeling previously described by us during P19SCs differentiation [6]. CL has emerged as an essential lipid in controlling the steps that lead to the release of cytochrome c to the intermembrane space [32]. In this context, CL peroxidation has been shown to lessen the binding of cytochrome c to the inner mitochondrial membrane and facilitate permeabilization of the outer membrane [33]. In fact, it was previously suggested that CL regulates cell death and survival depending on its level of peroxidation. Thus, peroxidized CL, that is present on the outer mitochondrial membrane, is able to initiate apoptosis, whereas non-peroxidized CL initiates mitophagy to improve mitochondrial quality avoiding apoptosis [34]. Then, our results showing a higher palmitoleic and linoleic acids, together with peroxidized CL in P19dCs, suggest a different mitochondrial redox status generated by the RA-induced differentiation that probably make these cells more prone to mitochondrial pathways of cell death. Contrarily, the integration of oleic acid into mitochondrial CL in P19SCs probably led to the desensitization of apoptosis via suppression of CL oxidation, as described in mouse embryonic stem cells [35].

Apoptotic machinery in SCs often shows relevant alterations that may play a role in the production of mitogenic signals [36]. In this context, the TRP53 accumulation was also reported in other types of ESCs in an inactive state, predominantly located in the cytoplasm [37,38] and constituting a driving force for ESCs differentiation by suppressing NANOG and POU5F1 expression throughout its transcriptional activities [39,40]. Our data suggest that differentiation of P19SCs activates TRP53, making differentiated cells competent to trigger efficient TRP53-mediated responses such as cell-cycle arrest and apoptosis. Moreover, the higher expression of CASP3 without being associated to the cleavage of common substrates in P19SCs supports the hypothesis about the role of TRP53 in SCs. In fact, it was also described that CASP3 is involved in embryonic development and tissue regeneration by stimulating the proliferation of stem and progenitor cells [41,42]. Thus, it was described that caspases play critical roles in induction of induced pluripotent stem cells from human fibroblasts [43]. At the time, it was described that P19 cells undergo apoptosis during neuronal differentiation induced by retinoic acid [44] and that caspase inhibition decreases the ability of P19 cells to differentiate into cardiomyocytes [45]. Altogether, these results suggest that the main components of the apoptotic pathway such as caspases and TRP53 present apoptosis-independent functions in SCs probably playing a role in maintaining an embryonic signature, regulating SCs differentiation and increasing the apoptotic threshold, conferring particular resistance against classical inducers of apoptosis.

Calpain activation in P19dCs points to intracellular calcium overload and participation of both ER and mitochondria. The ER has the ability to adapt to periodic cycles associated with metabolic demands of limited duration such as nutrient deprivation, hypoxia or changes in glycosylation status [46]. It was postulated that ER stress participates in ESCs differentiation through ER-associated protein degradation such as the UPS [47]. Therefore, our results showing the P19SCs differentiation-related activation of ATF6 and EIF2S1 suggest the triggering of ER stress responses associated with global protein synthesis and the improvement of protein folding which are probably related to the metabolic remodeling towards a more oxidative metabolism previously described by us [6]. In fact, it was described a role for EIF2S1 in promoting cell survival in response to oxidative stress. Thus, attenuation of protein synthesis by EIF2S1 probably prevents P19dCs from excessive proliferation, ATP depletion, stimulation of mitochondrial oxidative phosphorylation and ROS production [48].

When ER stress is prolonged and the ER is unable to recover its function, apoptosis is usually activated by inducing DDIT3 transcription to remove damaged cells. Although DDIT3 transcription is induced by the three UPR sensors, the EIF2S1 pathway that was activated in P19dCs plays a dominant role. Despite this, P19dCs presented lower DDIT3 levels. It was recently suggested that DDIT3 frames the conundrum of a pro-survival pathway that kills cells [49]. Accordingly, the higher level of DDIT3 in P19SCs probably constitutes an adaptive response to maintain clonal integrity by clearance of individual SCs after stressful events to prevent propagation of damaged SCs [50]. In addition, it was also described that EIF2AK3- and ATF6-deficient cells like P19SCs activate survival autophagy to counteract ER stress [51]. This particular adaptation of UPR probably contributes to select and maintain the SCs pool with the required properties and quality. Therefore, the higher expression of ERN1 and DDIT3 observed in P19SCs indicates an alteration in ER function in SCs and probably has additional meanings such as the upregulation of autophagy [52]. Furthermore, as ERN1 plays a function in mammalian developmental signaling [53], it may play a role in SCs maintenance and metabolic quiescence [54]. On the other hand, its depletion in P19dCs probably contributes to the protection of those cells from ER stress-induced cell death. Overall, this indicates that ER stress accompanies SCs differentiation and that activation of the ATF6 and EIF2S1 branches of UPR probably facilitates adaptation to the stress induced by the accumulation of unfolded/misfolded proteins targeted for degradation by UPS and CMA.

CTSD is considered a mediator of caspases activation [55], and CTSB:CTSD activity ratio is often used as indicator of lysosomal viability. Thus, our results suggest that P19SCs differentiation is accompanied by a controlled expansion of the lysosomal compartment, which may facilitate cathepsin activity without triggering lysosomal cell death processes. This greater lysosomal activity and the upregulation of UPS and CMA found in P19dCs would suggest the activation of additional systems of protein quality control, such as macroautophagy, to adapt SCs during differentiation and counteract the associated ER stress. Macroautophagy is a catabolic process that plays a role in the maintenance of SCs phenotype by controlling protein turnover via MTOR [56]. The activation of the PtdIns3K/AKT1/MTOR/RPS6KB1 pathway during SCs differentiation probably suppresses apoptosis, promotes cell growth, and drives cellular proliferation in response to nutrients, growth factors and cellular energy, through the modulation of translation, transcription, ribosome biogenesis, nutrient transport, and, as mentioned, autophagy. Therefore, this activation may be important in regulating quiescence, proliferation, differentiation potential and longevity of SCs being probably associated with metabolic reprograming towards a mitochondrial metabolism that produces higher amounts of ATP to meet the energetic requirements of differentiated progenies.

The relationship between oxidative stress, mitochondrial function, ER-stress and protein quality control mechanisms is dependent on the cell type [57], not being completely understood in SCs. Our data suggest different mechanisms to maintain protein homeostasis between undifferentiated and differentiated cells. Our results reveal that while protein quality control in P19SCs is mainly managed by macroautophagy, the UPR, UPS, chaperones and proteases exert a prominent role in their differentiated counterparts. Macroautophagy activation in P19SCs probably inhibits ER-stress, maintains viability of quiescent mitochondria, inhibits apoptosis, and retains cells in an undifferentiated state.

Although the inherent complexity of PtdIns3K/AKT/MTOR axis, their multiple branches and downstream effects constitute an obstacle to elucidate the concrete molecular details involved during early embryonic development, our results contribute to improve our understanding about the role of PtdIns3K, MTOR and autophagy in SCs and early differentiated cells. PtdIns3K has been extensively studied in the context of cancer and autophagy. However, its importance in development is understated, even though loss any of the core components of PtdIns3K often results in embryonic lethality [58,59]. Recent studies have begun to explore the contribution of PtdIns3K in ESCs maintenance [60]. It has become apparent that this signaling pathway has critical roles for the self-renewal of SCs, as confirmed by us in P19SCs when treated with 3-MA, which lost their active proliferation rate that is essential for self-renewal. In addition, it was described that PtdIns3K activation in ESCs may often occur through non-canonical means [60] and that its deletion or inhibition compromises mouse and human ESCs pluripotency and differentiation potential [61–63], as was also observed by us in P19SCs treated with wortmannin. Then, these results suggest that PtdIns3K support the retention of ESCs properties. On the other hand, P19SCs differentiation activated AKT1 and MTOR coupling cell growth and proliferation with the changes induced in energy metabolism upon differentiation. We have previously demonstrated that cellular metabolism plays an essential role in regulation of pluripotency in P19SCs that present a preference for a glycolytic metabolism when compared with their early-differentiated counterparts, which have high glycolytic rates but also utilize oxidative phosphorylation [6]. Indeed, and in agreement with our results, it was described that PtdIns3K plays a major role in regulating the cellular uptake of glucose and in stimulating several glycolytic enzymes, such as hexokinase [64], while MTOR regulates both glycolysis and oxidative phosphorylation and negatively regulates autophagy [65]. Furthermore, it has also been demonstrated that MTOR is required for the maturation and differentiation of multiple lineages [66,67]. Accordingly, the treatment with MTOR inhibitor rapamycin decreased the proliferation rate of early-differentiated P19 cells. In fact, although there are no severe developmental abnormalities associated with the deficiency of effectors downstream of MTOR, deletion of key components of MTOR leads to lethal phenotypes, highlighting the importance of MTOR for cell differentiation [60]. Although P19SCs present lower MTOR activation and higher autophagic flux than P19dCs, the inhibition of the autophagy machinery by using siRNA targeting Atg7 had not detrimental effects on cell proliferation, as also observed by other authors in human and murine ESCs under culture conditions with an abundance of nutrients [68,69]. Conversely, early-differentiated P19 cells require an intact autophagy machinery to sustain their proliferation. Despite this, Atg7 siRNA compromised the pluripotent state of P19SCs similarly to wortmannin confirming that deficiencies in autophagy machinery in ESCs leads to detrimental accumulation of pluripotency-associated proteins [69]. Consequently, reduced MTOR signaling and increased autophagy seem to be required to support ESCs homeostasis via the expression of pluripotency genes [68,69]. Altogether, these results unveil the new roles of PtdIns3K, AKT, MTOR and autophagy in the control of pluripotency, self-renewal, differentiation and cell fate determination during embryonic development.

P19SCs have weakly polarized, round and small mitochondrial bodies which became filamentous and fully polarized after differentiation. In addition, we detected in P19dCs an upregulation of mitochondrial biogenesis, function and ROS production [6], together with a promotion of mitochondrial fusion. Autophagic degradation of dysfunctional mitochondria is considered as a component of quality control mechanisms and the current models propose that fusion-/bioenergetic-incompetent mitochondria are targeted for mitophagy by increased free radical production [70]. We analyzed SQSTM1 recruitment to mitochondria in both types of p19 cells and detected a higher mitophagic flux in P19SCs, suggesting a higher mitochondrial quality/clearance in undifferentiated cells and the accumulation of fused and stressed mitochondria in P19dCs. Thence, the differential autophagy-dependent mitochondrial turnover in stem and differentiated cells probably contributes to maintain a mitochondrial phenotype that drives stemness [6] and is a key element in establishing the different susceptibility to cell death.

In summary, by playing a significant role in eliminating damaged macromolecules and preserving tissue homeostasis in the SCs pool, autophagy and apoptosis are coordinated in a way that avoids loss of quiescence, increases resistance to stress and damage, and minimizes the transmission of damaged structures to progeny cells. Thus, PtdIns3K and caspases inhibition affects P19SCs fate by losing pluripotency, pushing P19SCs to differentiate away from an endoderm fate and toward a neuronal fate, and impairing their ability to differentiate properly when treated with RA. On the other hand, autophagy inhibition through Atg7 silencing also induces a loss of pluripotency in P19SCs disturbing their ability to differentiate but without forcing a spontaneous neuronal differentiation. Then, our results demonstrate a key role for PtdIns3K/AKT/MTOR in pluripotency and cell fate determination and that pluripotency of P19SCs is associated with a high antioxidant protection, a deregulation in the apoptotic machinery and an activation of proteolytic systems that probably contribute to preserve cell quality, increasing the threshold for cell death activation.

Materials and methods

Cell culture and differentiation

P19 embryonal carcinoma cells were obtained from the American Type Culture Collection (CRL-1825) and cultured in glucose- or galactose (glucose-free)-containing media at 37 °C in a 5% CO2 atmosphere. High-glucose Dulbecco’s modified Eagle’s medium (DMEM; Sigma, D5648) was supplemented with 10% FBS (Gibco, 10270-106), 1.8 g/l sodium bicarbonate (Sigma, S5761), 110 mg/l sodium pyruvate (Sigma, P5280) and antibiotic/antimycotic solution (Sigma, A5955). Galactose (glucose-free)-containing medium was prepared using DMEM without glucose (Sigma, D5030) supplemented with 10% FBS, 1.8 g/l sodium bicarbonate, 110 mg/l sodium pyruvate, 1.8 g/l galactose (Sigma, G5388), 0.584 g/l l-glutamine (Sigma, G3126) and antibiotic-antimycotic solution. P19 stem cells were maintained in an undifferentiated state by growing them in monolayer and passaged every 2 days at a 1:20 to 1:30 dilution. To initiate neuroectodermal differentiation, P19 cells were seeded at a density of 5.2 × 10³ cells/cm² and treated with a single dose of 1 μM retinoic acid (Sigma, R2625). All experiments were performed after 4 days differentiation.

Cytochemistry and immunocytochemistry

To measure intracellular reactive oxygen species, P19 stem cells (P19SCs) and retinoic acid-differentiated cells (P19dCs) seeded on glass coverslips were incubated with the fluorescent probe CM-H₂DCFDA (7.5 μM; Invitrogen, C6827) for 30 min at 37°C in Krebs medium (1.8 mM CaCl₂, 132 mM NaCl, 4 mM KCl, 1.2 mM Na₂HPO₄, 1.4 mM MgCl₂, 6 mM glucose, 10 mM HEPES, pH 7.4), supplemented with 10% FBS. After 30 min incubation time, the media was replaced by fresh BSA-supplemented Krebs media. Coverslips were immediately imaged on an epifluorescence Leica DM 4000B microscope. Quantitative analysis of fluorescence intensity was measured by ImageJ software.

For labeling lysosomal bodies, P19 cells were seeded at the concentration described in the section on cell culture and differentiation above, in μ-Slide 8 well ibiTreat (ibidi, 80826). At the day of the experiment, growth medium was removed and cells were incubated with 75 nM LysoTracker Green (Invitrogen, L7526), a weakly basic amine that selectively accumulates in cellular compartments with low internal pH, in pre-warmed growth medium for 45 min under growth conditions. After incubation period, medium was replaced by fresh growth medium. Cells were examined by confocal microscopy using Zeiss LSM510 META and image analysis performed in the LSM 510 software.

For immunocytochemical analysis, cells were incubated with 125 nM MTR CMXRos (Invitrogen, M7512) in DMEM to label mitochondria. After 20 min at 37°C, the staining solution was replaced with fresh pre-warmed medium. Then, the incubation medium was removed and cells were fixed in 4% formaldehyde in PBS during 15 min at 37°C. After washing three times with PBS for 5 min, cells were permeabilized with 0.2% Triton X-100 (Sigma, X100) in PBS for 10 min. Cells were washed again three times in PBS, incubated in PBS with 1% BSA (Sigma, 1.12018) during 1 h at 4°C, probed with specific primary antibody against SQSTM1 (MBL International, PM045) in PBS with 1% BSA for 2 h at 37°C and incubated with a dilution (1:1,000 in PBS with 1% BSA) of a corresponding fluorescence-conjugated secondary antibody for 1 h at 37°C. Between the labeling with primary and secondary antibodies, cells were rinsed three times with PBS for 5 min. After labeling, coverslips were mounted on glass slides in Prolong Gold antifade medium (Invitrogen, P36934) and cells were imaged by confocal microscopy (Zeiss LSM 510Meta, Germany).

Determination of reactive nitrogen species, antioxidant defense, reduced/oxidized glutathione, protein carbonyl and lipid peroxidation

Nitrite concentration was measured as indicator of reactive nitrogen species production using the Griess reagent [71]. Briefly, cell culture extracts (100 μl) were added to 100 μl Griess reagent (1% sulfanilamide and 0.1% naphthylenediamine in 5% phosphoric acid). After 20 min incubation at RT, the optical density was measured at 540 nm in a microplate reader (BioTek PowerWave, Winooski, VT, USA). The concentrations of nitrite were calculated by comparison of absorbance of standard solutions of sodium nitrite prepared in culture medium and assayed under similar conditions.

The antioxidant defense was analyzed in cell extracts by measuring total antioxidant activity with the ABTS cation radical method [72] and catalase and SOD activities, according to methods already described [73,74]. Reduced and oxidized glutathione (GSH and GSSG) were evaluated by HPLC (Gilson, Lewis Center, Ohio, USA) with fluorimetric detection (excitation at 385 nm and emission at 515 nm; FP-2020/2025, Jasco, Tokyo, Japan), using the Immunodiagnostik kit (Immunodiagnostik AG, KC1800).

Protein carbonyl concentrations were measured in extracts from both types of P19 cells following the method developed by Levine [75] with modifications of Coto-Montes and Hardeland [76], as indicator of protein oxidative damage.

Lipid peroxidation end-products malondialdehyde (MDA) and 4-hydroxynonenal (4-HNE) were determined using the 1-methyl-2-phenylindole method, based on the condensation of the chromogenic agent 1-methyl-2-phenylindole with either MDA or 4-HNE [77].

Western blot analysis

In order to obtain total cellular extracts, P19 cells were harvested by trypsinization, washed with PBS and centrifuged for 5 min at 1,000 × g. The cellular pellet was resuspended in RIPA buffer (Sigma, R0278) supplemented with 2 mM dithiothreitol, 100 μM phenylmethylsulfonyl fluoride) and a protease inhibitor cocktail (containing 1 μg/ml of leupeptin (Roche, 11017101001), antipain (Roche, 11004646001), chymostatin (Roche, 11004638001) and pepstatin A (Roche, 10253286001)), physically ruptured by sonication and kept at −80°C until used. Protein contents were determined by using the BCA protein assay (Thermo Scientific, 23227). After denaturation at 95°C for 5 min in a Laemmli buffer (Bio-Rad, 161-0737), equivalent amounts of protein (50 μg) were separated by electrophoresis in 8, 12 or 14% SDS-polyacrylamide gel and electrophoretically transferred to a polyvinylidene difluoride membrane. After blocking membranes with 5% skim milk in TBS-T (50 mM Tris-HCl, pH 8, 154 mM NaCl and 0.1% Tween-20) for 1 h at room temperature, membranes were incubated overnight at 4°C with the antibodies directed against SOD2 (sc-18504), NFE2L2 (sc-722), TAZ/tafazzin (sc-49760), ATF6 (sc-22799), TRP53 (sc6243), SPTAN1 (sc-46696), CTSD/cathepsin D (sc-6486), MTOR (sc-8319), p-MTOR (S2448) (sc-101738), DNM1L (sc-32898), OPA1 (sc-30573), TOMM20 (sc-11415), TUBB3/betaIII-tubulin (sc-80005) from Santa Cruz Biotechnology; p-EIF2S1 (S51) (3398), ERN1 (3294), DDIT3 (2895), ubiquitin (3933), BCL2 (2870), BAX (2772), BBC3 (7467), Cleaved CASP3 (9661), AKT (9272), p-AKT1 (S473) (9271), RPS6KB1 (9202), BECN1/Beclin-1 (3495), ATG3 (3415), ATG12 (4180), POU5F1 (2840), SOX2 (2748) from Cell Signaling Technology; LAMP2A (Ab18528), NANOG (ab80892) from Abcam; MAP1LC3 (PD014), SQSTM1 (PM045) from MBL International; SHC1 (BD Biosciences, 610879), p-SHC1-1/p66SHC (S36) (Calbiochem, 566807), CRLS1 (Biorbyt, orb35719), BNIP3L (Sigma, N0399), DRAM1 (Rockland, 600-401-A70) and KRT8 (Developmental Studies Hybridoma Bank, TROMA-I-s), each previously diluted 1:1,000 in blocking buffer (1% skim milk in TBS-T). After 3 5-min washes in PBS-T, the membranes were incubated with a dilution (1:10,000 in blocking buffer) of a corresponding anti-rabbit (Cell Signaling Technology, 7074), anti-goat (Santa Cruz Biotechnology, sc-2020) or anti-mouse (Cell Signaling Technology, 7076) IgG horseradish peroxidase-conjugated secondary antibody for 2 h at room temperature. After three 15-min washes in PBS-T, membranes were developed with the ECL detection system (Millipore, WBKLS0500) and imaged with Versa Doc imaging system (Bio-Rad) according to the manufacturers’ protocols. Densities of each band were calculated with Quantity One Software (Bio-Rad). All data presented are representative from three separate experiments. Since RA-induced differentiation alters housekeeping proteins related with energy metabolism and cytoskeleton [6], Ponceau S staining was used to normalize western blot quantifications and ensure equal protein loading, following suggestions from the literature [78,79].

Lipid extraction and separation

P19 cells were washed with ice-cold PBS, scraped into 5 ml ice-cold PBS and centrifuged at 200 g for 4 min. The final cell pellet was resuspended in 1 ml dH₂O. Total lipids were extracted by the Bligh and Dyer method [80]. Briefly, 3.75 ml chloroform-methanol 1:2 (v/v) and a low amount of BHT (5 µg/ml) were added to the suspension that was then vortexed and incubated on ice for 30 min. Finally, 1.25 ml chloroform and 1.25 ml dH₂O were added. After vigorous vortex-mixing, the samples were centrifuged at 1,000 g for 5 min at RT to obtain a two-phase system: an aqueous top phase and an organic bottom phase from which lipids were obtained. The total lipid extracts, recovered from the bottom phase, were dried under N₂ gas and stored at −20°C.

Total lipid extracts were separated into different classes by thin-layer chromatography (TLC). Several spots corresponding to 30 μg total lipid extract dissolved in chloroform were applied to a TLC plate with a concentrating zone 20 cm × 20 cm, 0.65 mm (Merck, Darmstadt, Germany). Before separation, plates were washed with chloroform-methanol and, after drying, sprayed with 2.3% boric acid in ethanol and exposed to high temperature (100°C for 15 min). The plates were developed with chloroform–ethanol–water–triethylamine 30:35:7:35 (v/v) as mobile phase. Lipid spots on the silica plates were observed by spraying the plates with primuline (5 mg/100 ml) and identified (UV lamp at 260 nm) by comparison with authentic lipid standards. Then, classes from five spots were scraped off the plates: one spot was quantified with the phosphorus assay, and four spots were extracted with the use of chloroform/methanol (2:1) (v/v) for subsequent identification by mass spectrometry (MS).

Phospholipid quantification

Phospholipids (PL) in lipid extracts or in chromatographic fractions were estimated by phosphorus determination through an acidic digestion. The released inorganic phosphate was reacted with ammonium molybdate, with the complex giving a strong blue color. To evaluate the phospholipid content, phosphorus assay was performed by a method described elsewhere [81]. Briefly, 500 µl perchloric acid (70%) was added to phosphate standards of KH₂PO₄ (1–5 µg) and samples (25 µl; 50 µl and 100 µl) previously dried. The samples were incubated for 1 h at 180°C, followed by cooling to room temperature. Then, 3.3 ml of water, 0.3 ml ammonium molybdate, and 0.3 ml ascorbic acid were added to the standards and samples, which were then incubated for 10 min at 100°C in a water bath. The classes from spots scraped to silica plates were centrifuged for 5 min at 1,000 × g to separate phospholipids from silica. Finally, the standards and samples solutions were measured at 800 nm. The relative abundance of each PL class was calculated by relating the amount of PL class with the total amount of PL in the sample.

Electrospray mass spectrometry

Analysis of cardiolipin (CL) was carried out in negative mode on electrospray ionization (ESI) LXQ linear ion trap mass spectrometer (ThermoFinnigan, San Jose, CA, USA). The samples for electrospray analysis were prepared by diluting in 20 µl CL solution (obtained after extraction from silica and re-suspended in 100 µl of chloroform) with 100 µl methanol. The samples were introduced into the mass spectrometer using a flow rate of 8 µl/min. ESI conditions in electrospray linear ion trap mass spectrometer were also as follows: voltage, 4.7 kV; capillary voltage −34.9 V; tube lens voltage, −124.9 V; capillary temperature, 275°C; and the sheath gas flow was 25 units. Isolation with an interval of 0.5 Da was used with a 30 milliseconds activation time for MS/MS experiments. Full scan MS spectra and MS/MS spectra were acquired with 50 and 200 ms maximum ionization time. Normalized collision energy (CE) was varied between 17 and 20 (arbitrary units) for MS/MS. Data acquisition of this mass spectrometer was carried out using an Xcalibur data system (V2.0).

Cardiolipin peroxidation

CL peroxidation was measured using the 10-N-nonyl acridine Orange (NAO) fluorescent probe. NAO is able to bind with high affinity to non-oxidized cardiolipin (green-fluorescence). When in the presence of oxidized CL, NAO has been reported to bind to CL with a decrease affinity [82]. Thus, we analyzed the loss of the fluorescence peak of NAO that is proportional to CL peroxidation levels. P19 cells were seeded in 100 mm plates at a concentration of 0.5 × 10⁴ cells/ml for P19SCs and 2 × 10⁴ cells/ml for P19dCs. Twenty-four hours after seeding both groups of cells were also treated with 500 µM H₂O₂ during 4 h as a positive control. The cells were than washed, trypsinized and centrifuged at 210 × g for 10 min at 4°C. Cell pellets were then washed in 1 ml PBS and centrifuged again. The resultant cell pellets were re-suspended in 0.1 µM NAO in PBS, incubated for 30 min at 37°C/5% CO₂ in darkness. Then, cells were centrifuged at 210 × g for 5 min at 4°C and the cell pellet was re-suspended in 0.3 ml of PBS and transferred to FACS tubes. The cell suspension (20,000 cells) was analyzed by FACSCalibur flow cytometer. The fluorescence excitation of NAO was measured at 488 nm while the fluorescence emission at 530 ± 15 nm. Two regions were defined, M1 and M2, based on the fluorescence of P19SCs that presented the highest NAO fluorescence (M2 region) and, consequently, the lowest levels of peroxidized CL. The appearance of a peak in M1 region towards loss of fluorescence in M2 region indicated an increase of CL peroxidation. All experimental conditions were compared based on those regions.

Proteasome activity

Proteasome activity was measured with the 20S Proteasome Activity Assay Kit (Merck Millipore, APT280). Briefly, P19 cells were collected as described before, re-suspended in RIPA buffer supplemented with 1 mM DTT and, after protein quantification, cell extracts were stored at −80ºC. In a 96-well plate, cells were incubated with 80 µl of Assay buffer (250 mM HEPES, pH 7.5, 5 mM EDTA, 0.5% NP-40, 0.01% SDS) and with 10 µl proteasome substrate LLVY-7-amino-4-methylcoumarin (AMC). Twenty-five µl of the proteasome inhibitor lactacystin were added to some samples. As a positive control, 10 µl of 20S Proteasome Positive Control (25 μl 20S Proteasome in 50 mM HEPES, pH 7.6, 150 mM NaCl, 1 mM DTT) were also added to the plate. After 1–2 hours incubation at 37°C, the fluorescence was read in a microplate reader (BioTek PowerWave) using a 380/460 nm filter. The standard curve ranging from 0.04 μM – 12.5 μM was made by using known solutions of AMC.

Caspase-like activities

Caspases activities were accessed using colorimetric assays based on the proteolytic cleavage of a substrate composed by the chromophore, p-nitroanilide (pNA) and a synthetic tetrapeptide. Cells extracts were trypsinized and washed in PBS. After washing, cells were concentrated by centrifuging at 300 × g for 3 min and the supernatant was discarded. The pellets were re-suspended in a lysis buffer containing 100 mM NaCl, 0.1% CHAPS, 1 mM DTT, 0.1 mM EDTA, 50 mM HEPES pH 7.4. The samples were kept on ice for 20 min following by protein quantification using the BCA kit assay (Thermo Scientific, 23227). Then, 25 μg of protein extracts were aliquoted in an assay buffer (100 mM NaCl, 0.1% CHAPS, 1 mM DTT, 0.1 mM EDTA, 10% glycerol, 50 mM HEPES pH 7.4). The assay buffer was supplemented with the corresponding substrate: 100 μM Ac-DEVD-pNA (Merck Millipore, 235400), 200 µM Ac-LEHD-pNA (Merck Millipore, 218805) or 200 µM Ac-IETD-pNA (BioVision, 1063) for measuring CASP3, CASP9 and CASP8 activities, respectively. Absorbance was measured at 405 nm in VICTOR X3 plate reader (Perkin Elmer) after 2 h incubation at 37°C. Caspase-like activity was expressed as pmol of pNA/μg of protein.

Live/dead assay

Calcein-AM (Invitrogen, C1430) and propidium iodide (Sigma Aldrich, P4864) were used to determine the percentages of viable and dead cells in P19SCs and P19dCs. 10⁶ cells were harvested, washed, resuspended in HBSS/Ca and loaded with 0.1 μM calcein-AM and 8 μM propidium iodide for 20 min at room temperature. Fluorescence was measured by FACS (Becton Dickinson FACScalibur) with 488 nm excitation wavelength. The simultaneous measurement of calcein/propidium iodide fluorescence was performed using 530/30 nm bandpass filter for calcein and a 610/20 nm bandpass filter for propidium iodide red fluorescence.

CTSD/cathepsin D and CTSB/cathepsin B activities

CTSB and CTSD activities were measured by fluorimetric and spectrophotometric detection, respectively, as previously described [83]. P19 cells were collected as mentioned above and resuspended in a Lysis Buffer. The samples were kept on ice for 20 min following by protein quantification through the BCA assay. To measure the CTSB-like activity, the cells were incubated with 40 μM of Z-Arg-Arg-N-methyl-coumarin (Sigma, C5429) in 100 mM sodium acetate pH 5.5, 1 mM EDTA, 5 mM DTT and 0.1% Brij-35 at 37ºC for 20 min. After the incubation, 150 μL of stopping buffer (33 mM sodium acetate pH 4.3 and 33 mM sodium chloroacetate) were used to stop the enzymatic reaction. CTSB activity were determined by the detection of the N-methyl-coumarin (Sigma, A9891) fluorometrically at 360 nm excitation and 460 nm emission in a VICTOR X3 reader (Perkin – Elmer, Boston, MA, USA). The method was calibrated with known concentrations of N-methyl-coumarin. For CTSD activity determination, the cells were incubated with 125 μL of 3% hemoglobin (Sigma, H2625) in 200 mM acetic acid at 3°C for 30 min. After the incubation, 125 μL of 15% TCA were added to the samples and they were kept at 4°C for 30 min. Samples were then centrifuged at 13,400 × g for 5 min and CTSD activity was determined by the measurement of 200 μL supernatant at 280 nm in a Cytation 3 (BioTek Instruments) multi-plate reader.

Electron microscopy

Electron microscopy was performed to examine the complexity of the endo-lysosomal system in both P19SCs and P19dCs. Cells were trypsinized and fixed for electron microscopy in 3% glutaraldehyde in cacodylate buffer (0.1 M sodium cacodylate buffer), pH 7.3 for 2 h at 4°C. Cells were then centrifuged for 3 min at 2500 × g, washed with cacodylate buffer and incubated in 1% OsO4 in cacodylate buffer for 2 h. Pellets were again washed in cacodylate buffer and embedded in 1% agar. Samples were then dehydrated in ethanol and embedded in Spurr’s resin. An LKB ultra-microtome Ultrotome III (GE Healthcare, Buckinhamshire, UK) was used to obtain thin sections that were subsequently stained with methanolic uranyl acetate followed by lead citrate. Sections were examined with a JEOL Jem-100SX electron microscope (JEOL, Tokyo, Japan) operated at 80 kV. Forty micrographs of each sample were taken to examine the endo-lysosomal system in both groups of cells.

Isolation of mitochondrial extracts

Mitochondrial extracts were isolated harvesting P19 cells by trypsinization and spinning them down at 1,000 × g. Pellets were washed once in cold PBS and centrifuged again (1,000 × g) at 4°C. The cell suspension was then resuspended in 0.5 ml of ice-cold sucrose buffer (250 mM sucrose, 20 mM HEPES pH7.5, 10 mM KCl, 1,5 mM MgCl2, 0,1 mM EDTA, 1 mM EGTA) supplemented right before use with 1 mM DTT, 0.1 mM PMSF and protease inhibitor cocktail containing 1 mg/ml of leupeptin, antipain, chymostatin and pepstatin. The suspension was then incubated on ice for 20–30 min. After incubation, cells were transferred to a pre-cooled tissue homogenizer and homogenized 30 times using a tight pestle, while keeping the homogenizer on ice. The progress was monitored every 20–30 strokes under a phase contrast microscope and was stopped when more than 90% of cells were burst. Homogenized cells were centrifuged at 3500 × g for 5 min at 4°C. The supernatant was collected containing mitochondrial and cytosolic fraction. Then, the collected supernatant was centrifuged again at 10,000 × g during 15 min at 4°C. The pellet, corresponding to the mitochondrial fraction was resuspended in 50 µl of sucrose buffer cited above.

Autophagic flux determination

To measure autophagic flux, the fusion of autophagosomes with lysosomes was blocked by treating cells with bafilomycin (Cayman, 11038). Cells were seeded at 2 × 10⁴ cell/ml for P19SCs and 5 × 10³ for P19dCs and twenty-four hours after seeding, they were treated during 24 h with 0.2 µM bafilomycin. To determinate autophagic flux MAP1LC3-II (MBL, PD014) was detected in whole-cell extracts and to measure mitophagic flux SQSTM1 (MBL, PM045) and TOMM20 (translocase of outer mitochondrial membrane 20, Santa Cruz Biotechnology, sc-11415) were detected in mitochondrial extracts, by western blotting as described above.

Cell transfection

Cells were transfected with 50 nM of either Atg7 siRNA oligonucleotide (SI00900529, Qiagen; Hilden, Germany) or with a scrambled siRNA (D-001810-03-20; Dharmacon, Bucks, UK). To perform the assay, 24 h before the transfection, P19 cells were seeded at a density of 7.5 × 10³ cell/cm². Transient transfection was performed using Lipofectamine 2000 transfection reagent (Invitrogen, Paisley, UK) according to the manufacturer’s instructions. In order to verify whether Atg7 silencing influences differentiation capacity, P19 cells were subjected to the differentiation protocol 24 h after transfection was performed.

Cell proliferation assay

Cell proliferation was measured by the sulforhodamine B (SRB) assay as described previously [84–86]. P19 cells were seeded in 48-well plates at a concentration of 5 × 10³ cells/ml for P19SCs and 2 × 10⁴ cells/ml for P19dCs. Twenty-four hours after seeding, specific wells were treated with the oxidative stress inducer H₂O₂ (500 µM and 1 mM), the MTOR inhibitor rapamycin (0.1 nM, 10 nM and 100 nM), and the PtdIns3K inhibitor 3-MA (2,5 mM and 5 mM) to evaluate their effect on cell proliferation. Both types of cells were also transfected with either Atg7 siRNA oligonucleotide (si-Atg7) or with a scrambled siRNA (si-Control). After 72 hours for treatments with 3-MA, rapamycin and siRNA, and after 2, 4, and 6 hours for treatments with H₂O₂, the medium was removed and wells rinsed with 1% PBS. Cells were then fixed by adding 1% acetic acid in 100% methanol for at least 2 h at −20°C. Later, the fixation solution was discarded and the plates dried in an oven at 37°C. Two hundred and fifty microliters of 0.5% SRB (Sigma, S9012) in 1% acetic acid solution were added and incubated at 37°C for 1 h. The wells were then washed with 1% acetic acid in water and dried. Then, 500 µl of Tris, pH 10 was added and the plates were shaken for 30 min. Finally, 200 µl of each supernatant was transferred in 96-well plates and optical density was measured at 540 nm.

Impact of PtdIns3K and apoptosis inhibition on cell differentiation

To evaluate the effect of apoptosis and PtdIns3K inhibition on stemness maintenance, one day after seeding, P19SCs were cultured for 24 h in the presence of the pan-caspase inhibitor Z-VAD-FMK (50 μM; Selleckchem, S7023) or with 100 nM wortmannin (Selleckchem, S2758), a PtdIns3K inhibitor. To evaluate the impact of both drugs on P19 cells differentiation, 24 hours after the treatment with RA, 50 μM Z-VAD-FMK or 100 nM wortmannin were added to growth medium and cells were harvested after the 96-hours period of differentiation discarding floating dead cells. Because wortmannin is unstable at 37°C in culture medium, it was added every 6 h to the cell culture. As controls, cells were incubated in the absence of WT for the same times. Markers of stemness (POU5F1, NANOG and SOX2) and differentiation (KRT8 and TUBB3) were detected using western blot according to the protocol described above.

Statistical analysis

Data are mean values ± S.D. calculated from at least three separate experiments. Statistical comparisons between P19SCs and P19dCs were carried out using two-tailed Student’s t-test. Multiple comparisons were performed using one-way or two-way ANOVA followed by the Bonferroni post-hoc test. Significance was accepted with p < 0.05.

Funding Statement

This work was supported by the European Regional Development Fund [POCI-01-0145-FEDER-016390:CANCEL STEM];European Regional Development Fund [POCI-01-0145-FEDER-007440];Fundação para a Ciência e a Tecnologia [IF/01316/2014];Instituto de Salud Carlos III [FI14/00405];Instituto de Salud Carlos III [PI17/02009];Instituto de Investigación Sanitaria del Principado de Asturias.

Acknowledgments

This work was supported by the European Regional Development Fund (ERDF) through the COMPETE 2020 - Operational Programme for Competitiveness and Internationalisation and Portuguese national funds via FCT – Fundação para a Ciência e a Tecnologia, under grants POCI-01-0145-FEDER-016390:CANCEL STEM, POCI-01-0145-FEDER-007440, and IF/01316/2014; the Spanish Ministry of Science, Innovation and Universities-ISCIII under grant PI17/02009. J.C. B-M. acknowledges his predoctoral fellow from the Instituto de Investigación Sanitaria del Principado de Asturias (ISPA). Y.P. acknowledges her predoctoral fellow (FI14/00405) from the Instituto de Salud Carlos III (Spanish Ministry of Science, Innovation and Universities).

Author contributions

S.M-N. and J.C.B-M performed the experiments and analyzed the data. R.L., K.A.M., J.R.E. and Y.P., helped in data acquisition and analysis. M.R.D., E.M. and T.M. helped in performing phospholipids experiments. I.B. helped in performing oxidative stress experiments. A.C-M. and J.H. provided suggestions and revised the manuscript. P.J.O. and I.V-N. conceptualized, designed and supervised the study. I.V-N. wrote the manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed here.