Mitochondrial biogenesis is not changed with NSC differentiation

The neurosphere assay, which was previously characterized by our group (Soares et al, 2020), is an appropriate model to evaluate the postnatal NSC fate as both stemness and multipotency properties of these cells can be determined in a heterogenic microenvironment that mimics the neurogenic regions. Notably, this model allows the simultaneous evaluation of postnatal differentiation into neurons, astrocytes, and oligodendrocytes (Weiss et al, 1996). To evaluate mitochondrial properties with NSC differentiation, NSCs were isolated from the SVZ. To guarantee a high yield of NSC population, we performed two passages, obtaining tertiary neurospheres. The cells within the neurospheres exhibit both stemness and proliferative capacity (Soares et al, 2020). SVZ neurospheres when plated under differentiative conditions (with the removal of growth factors from the medium) give rise to neuronal, oligodendroglial, and astroglial cells as shown in our previous work (Soares et al, 2020). In fact, throughout differentiation besides astrocytes, immature cells such as NSCs (SOX2+ cells), immature neurons (DCX+ cells), and oligodendrocyte precursor cells (OPCs) (PDGFRa+/NG2+ cells) are highly present at DIV2, whereas at DIV7, mature neurons (NeuN+ cells) and myelinating oligodendrocytes (MBP+ cells) are more expressed.

To investigate whether mitochondrial biogenesis is altered with postnatal NSC differentiation, we assessed the protein levels of the master regulators of this process: peroxisome proliferator–activated receptor-γ coactivator 1α (PGC1α) and the mitochondrial transcription factor A (TFAM). Once PGC1α is activated, this protein interacts with the nuclear respiration factors 1 and 2 (NRF-1 and NRF-2), leading to the expression of many mitochondrial genes and proteins that are needed for mitochondrial DNA replication and transcription, such as TFAM (Wu et al, 1999). Immunoblot analysis revealed no significant alteration in the protein levels of PGC1α (Figs 1A and B and S8A and B) and TFAM (Figs 1A and C and S8A and B) along NSC differentiation (DIV0, DIV2, DIV4, and DIV7). Therefore, our data suggest that mitochondrial biogenesis might not have an impact in SVZ-derived NSC differentiation.

NSC differentiation does not affect AMP-activated protein kinase (AMPK) protein levels

To further corroborate the previous findings indicating that mitochondrial biogenesis is not altered during SVZ-derived NSC differentiation, we explored the protein levels of indirect regulators of this process: AMPK and the mechanistic target of rapamycin (mTOR). AMPK is a heterotrimeric serine/threonine protein kinase with the capacity to phosphorylate PGC1α, through the reduction of ATP/AMP ratio in the cell (Bergeron et al, 2001; Reznick et al, 2007). Moreover, this kinase has a catalytic α-subunit and a scaffolding β subunit (β1 and β2). As for the master regulators, no changes were observed in the protein levels of AMPKα (Fig 2A and C), AMPKβ1 (Fig 2A and E), and AMPKβ2 (Fig 2A and G). In addition, no significant alterations were observed in phospho-AMPKα (p-AMPKα) and p-AMPKβ1, the respective activated forms of AMPKα and AMPKβ1 (Fig 2B, D, and F). Also, no significant differences were observed in the p-AMPKα/AMPKα (Fig S1A) and p-AMPKβ1/AMPKβ1 (Fig S1B) ratios. Hence, these results demonstrate that the overall protein levels of AMPK and its activated forms do not change throughout SVZ-derived NSC differentiation. Finally, mTOR another indirect regulator of mitochondrial biogenesis (Morita et al, 2013; Knobloch & Jessberger, 2017) also did not reveal significant changes in protein levels with NSC differentiation (Fig 2H). Overall, these findings further suggest that NSC differentiation does not affect mitochondrial biogenesis-related protein levels.

Mitofusins 1 and 2 increase, whereas DRP1 decreases, during NSC differentiation

To disclose whether NSC differentiation and maturation affect mitochondrial dynamics, the protein levels of the outer membrane fusion proteins Mitofusin 1 (Mfn1) and Mfn2 (Rojo et al, 2002) and the inner membrane fusion protein OPA1 (Cipolat et al, 2004) were evaluated in SVZ cells (Figure S9A-D. Interestingly, Mfn2 protein levels significantly increased with NSC differentiation, reaching a maximum level of 3.457 ± 0.3969-fold (n = 4–5, P < 0.001) in cells at DIV7 (Fig S2B). In contrast, no significant alterations were observed in Mfn1 (Fig S2A) and in the OPA1 full-length form (upper band) and cleaved form (lower band) (Fig S2C and D). To further understand the role of NSC commitment in mitochondrial dynamics, we also explored mitochondrial fission by evaluating the protein levels of the dynamin-related protein (DRP1) (Smirnova et al, 2001). Noteworthily, SVZ cell differentiation induced a marked decrease of DRP1 to 20% ± 2.85% of its initial levels (DIV7: 0.2018 ± 0.02854, n = 4–5, P < 0.0001) (Fig S2E).

To determine whether these differences observed in the protein levels of mitochondrial fusion–related proteins were dependent on alterations in mitochondrial mass during NSC differentiation, the mitochondrial mass was assessed by quantifying HSP60 protein levels, a mitochondrial matrix protein (Green & Kroemer, 2004). As HSP60 levels have no major change along cell differentiation (Figs 3 and S9), Mfn1/2 and OPA1 protein levels were normalized to HSP60. Notably, a significant increase in Mfn1 protein levels was observed in SVZ cells only at DIV7 (2.287 ± 0.3801, n = 4–5, P < 0.05) (Fig 3C). In respect to Mfn2 levels, the significant alterations were maintained (DIV2: 1.230 ± 0.1821, DIV7: 3.724 ± 1.161, n = 4–5, P < 0.05) (Fig 3D). No significant changes were observed in the OPA1 full-length and cleaved form (Fig 3E and F). Regarding DRP1, the fission-related protein continued to significantly decrease with cell differentiation reaching 25.10% ± 10.91% of the initial levels (DIV7: 0.2510 ± 0.1091, n = 4–5, P < 0.0001) (Fig 3G). Overall, our results suggest a role for mitochondrial fusion/fission machinery in SVZ-derived NSC differentiation.

Each neural cell type displays a unique mitochondrial morphology profile

Because the mitochondrial network is remodeled by fusion and fission events, and based on the results obtained in Fig 3, we postulated that both undifferentiated and differentiated cells could present distinct mitochondrial morphologies that can be lineage-dependent and maturation degree–related, in SVZ cells. To further explore this hypothesis, mitochondrial morphology was studied within undifferentiated (Fig S3) and differentiated cells (Figs 4, S4, and S5) including astrocytes, immature and mature neurons, OPCs, and different stages of oligodendrocytes. By using the MiNA macro (Valente et al, 2017), we evaluated several morphometric parameters including the number of individuals (rods and unbranched puncta), the number of networks (branched structures), the average number of branches per network, the average length of the rods and branches, and the percentage of mitochondrial area (Fig 4A). Importantly, these parameters, when considered together, allow to draw conclusions about mitochondrial fragmentation and fusion/fission events. Interestingly, the number of mitochondria, individuals, and networks significantly increased in NSCs (SOX2+ cells) from DIV0 to DIV2 (Fig S3B and C), during astroglial differentiation (GFAP+ cells) and in mature neurons (αtau+ cells) (Figs 4B and C and S4A and C). No significant changes were observed in immature neurons (DCX+ cells) from DIV2 to DIV4 regarding mitochondrial individuals (Figs 4B and S4B) and mitochondrial networks (Figs 4C and S4B), whereas in mature neurons, an increase is observed with differentiation. By contrast, no significant changes in the mitochondrial number were observed in OPCs (PDGFRα+/NG2+ cells) with differentiation (Figs 4B and C and S5A). Curiously, the number of mitochondria is lower in the more mature oligodendrocytes (MBP+ cells with complex branched structures) when compared with less mature (MBP+ cells with poorly branched structures) at both DIV4 and DIV7. However, the number of mitochondria was already low in mature oligodendrocytes at DIV4 and had similar values when comparing with mature oligodendrocytes at DIV7, suggesting that the number of mitochondria decreases to lower levels with oligodendrocyte maturation (Figs 4B and C and S5B). Interestingly, the more mature oligodendrocytes present a lower number of mitochondria when compared with astrocytes and neurons (Figs 4B and C, S4A–C, and S5B).

A significant reduction in the number of branches per network was observed in astrocytes from DIV4 to DIV7 and in both immature neurons and mature neurons (Figs 4D and S4A–C), whereas no alterations in NSCs (Fig S3D) and oligodendroglial cells (Figs 4D and S5A and B) were observed. Interestingly, in immature neurons, the mitochondria became more elongated as cells mature returning then to a less elongated phenotype in mature neurons (Figs 4E and S4B and C). In astrocytes, mitochondria become less elongated with differentiation but then return to a more elongated morphology (Figs 4E and S4A). NSCs revealed a significant decrease in the mitochondrial length from DIV0 to DIV2 (Fig S3E). Moreover, this parameter remained unchanged in oligodendroglial cells (Figs 4E and S5A and B). Lastly, when we analyze the mitochondria area in the cells, we found a significant increase in astrocytes (Figs 4F and S4A) along differentiation. In addition, as oligodendrocytes undergo maturation, the mitochondrial area is reduced to approximately half (Fig 4F) and a perinuclear distribution of the mitochondria is observed (Fig S5B). Therefore, these robust alterations of mitochondrial morphology could suggest that mitochondrial dynamics could be important modulators of SVZ-derived NSC fate; however, further validation is required.

NSC differentiation at the later stages requires higher levels of ATP production

Because mitochondrial dynamics and morphology are to a certain extent linked to mitochondrial bioenergetics, the dissimilarities observed in the different fates derived from NSCs might be because of differences in the bioenergetic profile. Therefore, respiratory assays were performed in control cells at different differentiation time-points (24, 96, and 96 h + 4 d in low EGF/FGF-2 medium). The extracellular acidification rate, a proxy for glycolysis, and the oxygen consumption rate (OCR) mediated by mitochondrial respiration were assessed by using the Mito Stress setup (Faria-Pereira et al, 2022). The energy map (Fig 5A) revealed that under stressed conditions (upon FCCP injection), these three controls do not present major changes. Nevertheless, we then evaluated the mitochondrial respiration rates (Fig S6A and B). No changes were observed in all the parameters (Fig 5B, D–G), with the exception of the ATP-linked respiration where a significant increase with differentiation and cell maturation from the 96 h to the 96 h + 4 d time-points (96 h: 66.30% ± 5.239% and 96 h + 4 d: 77.19% ± 2.925%, n = 6–7, P < 0.05) was observed (Fig 5C). Overall, these data suggest that throughout NSC differentiation, particularly at the later time-points, there was a more pronounced reduction in the % of OCR when mitochondrial ATP production was inhibited.

Neurons are the most energetically flexible cells differentiated from the SVZ-derived NSCs, whereas oligodendrocytes are the least flexible

We then assessed the energy map of the sorted astrocytes, oligodendrocytes, and neurons (Fig 6A), showing that the neurons are more aerobic when exposed to stress condition in comparison with glial cells that present a more glycolytic profile. In addition, neurons present significantly lower levels of basal respiration (astrocytes: 74.65% ± 1.616%, oligodendrocytes: 75.03% ± 2.347%, and neurons: 61.41% ± 4.538%, n = 4–9, P < 0.01) (Figs 6B and S7A) and proton leak (astrocytes: 25.35% ± 1.541%, oligodendrocytes: 24.49% ± 2.295%, and neurons: 15.05% ± 1.474%, n = 4–9, P < 0.01) (Fig 6D) in comparison with the glial cells. In addition to this, ATP-linked respiration did not change among the different cell types (Fig 6C). Neurons present a significantly higher maximal respiration in comparison with glial cells (astrocytes: 143.50% ± 6.526%, oligodendrocytes: 117.60% ± 9.755%, neurons: 184.80% ± 24.83%, n = 4–9; P < 0.05 and P < 0.01) (Figs 6E and S7B) and a significantly higher spare respiration in comparison with both astrocytes and oligodendrocytes (astrocytes: 68.82% ± 6.359%, oligodendrocytes: 42.61% ± 10.30%, neurons: 123.8% ± 26.31%, n = 4–9, P < 0.01 and P < 0.001) (Figs 6F and S7B), which suggests that cells that commit to the neuronal lineage present higher energy flexibility. Finally, significant differences were observed in non-mitochondrial respiration between neurons and glial cells (Fig 6G). Concluding, neurons are the most energetically flexible cells despite having the lowest basal respiration among the differentiated cells from the SVZ-derived NSCs.

Astrocytes have a significantly higher ATP content when compared with oligodendrocytes and neurons

To assess if the changes observed upon ATP-linked production (Fig 5D) in 96 h + 4 d control comparing with the 96 h were reflected in total ATP content, the levels of ATP were assessed in the 24, 96, and 96 h + 3 d controls and in the sorted cells (astrocytes, oligodendrocytes, and neurons).

No significant differences were observed in the ATP content levels when comparing all control conditions (Fig 7A). Importantly, the fact that the ATP content is similar between the 96 h and the 96 h + 3 d (Fig 7A) and because the 96 h + 4 d presents a significantly higher ATP-linked to respiration (Fig 5D) suggests that at 96 h + 3 d, there could be a higher rate of ATP consumption or that at 96 h, there could be a higher rate of ATP production by other pathways. Interestingly, although no differences were observed in the ATP-linked respiration in the sorted cells (Fig 6D), we observe that astrocytes present a significantly higher ATP content (astrocytes: 3.488 ± 1.042 μM/μg protein, oligodendrocytes: 0.9669 ± 0.3219 μM/μg protein, neurons: 0.4758 ± 0.2264 μM/μg protein, n = 4, P < 0.01 and P < 0.001) (Fig 7B). Overall, the data of the ATP content together with the ATP-linked respiration give new insights about the ATP consumption and the ATP production by other pathway in some conditions.

Animals

Mice were obtained from the iMM|JLA Rodent Facility (Lisbon, Portugal), where they were housed in a temperature-controlled room at 20–24°C. All the procedures were approved by the Portuguese National Authority for Animal Health (DGAV) and by the institute’s animals’ well-being office (ORBEA-iMM). This study was carried out in compliance with the ARRIVE guidelines (Percie du Sert et al, 2020). All experiments were performed in accordance with the European Community (86/609/EEC; 2010/63/EU; 2012/707/EU) and Portuguese (DL113/2013) legislation for the protection of animals used for scientific purposes.

In vitro cultures

SVZ NSCs were obtained from early postnatal (P1-3) C57BL/6 mice. This model is appropriate to mimic the physiological context of the postnatal NSC differentiation. Neurosphere culture was performed as previously described (Soares et al, 2020). SVZ were seeded at a density of 2 × 10⁴ cells/ml in serum-free medium (SFM) composed by DMEM/F12 + GlutaMAX-I supplemented with 100 U/ml penicillin and 100 μg/ml streptomycin, 1% B27, 10 ng/ml EGF, and 5 ng/ml bFGF (proliferation conditions) for 6–8 and 10–12 d, respectively. When most of the SVZ neurospheres have a diameter 150–200 μm, two passages were performed to obtain higher yields of NSCs, as described (Soares et al, 2020). SVZ neurospheres were plated (at a density of ∼60 neurospheres per well) onto glass coverslips coated with 100 μg/ml PDL in 24-well plates. For immunoblotting analysis, SVZ neurospheres (at a density of ∼480 neurospheres per well) were plated in coated six-well plates. The neurospheres were maintained in SFM devoid of growth factors (differentiation conditions). After 24 h, the medium was replaced with fresh SFM devoid of growth factors and at 4 and 10 d in vitro (DIV4 and 10), half of the medium was renewed. The SVZ-derived NSCs were allowed to develop for a maximum of 14 d (DIV14).

Immunoblot

Western blot analysis was performed to assess levels of proteins involved in mitochondrial dynamics and biogenesis during SVZ-derived NSC differentiation. SVZ neurospheres at P2 were plated and allowed to develop for 0, 2, 4, and 7 d in differentiation conditions. Cells were lysed with the lysis buffer composed by 1 mM EGTA, 250 mM sucrose, 5 mM Tris–HCl, and 1% Triton X-100, pH 7.4, supplemented with protease and phosphatase inhibitors. Protein concentration was measured by the Bradford method accordingly to the manufacturer’s specifications. 10–30 µm of proteins were separated by SDS–PAGE on 4–15% polyacrylamide gels (Bio-Rad) and electrophoretically transferred into a 0.2-μm nitrocellulose membranes. Membranes were blocked with 5% milk powder or 5% BSA in Tris-buffered saline with 0.1% Tween-20 for 1 h at RT. Incubations with the primary antibodies against proteins involved in mitochondrial dynamics were performed overnight at 4°C (Table 1). Secondary antibodies conjugated with the horseradish peroxidase enzyme were used and detected by the ECL chemiluminescent luminol substrate (Amersham) and imaged on the Amersham 680 equipment. To determine the AMPK activity, a ratio between the activated and total forms was calculated. GAPDH was chosen as a loading control because no changes were observed during NSC differentiation, as supported by other reports (Agostini et al, 2016; Zheng et al, 2016; Isaksen et al, 2020). HSP60 was also used as a loading control to normalize for the mitochondrial mass. In the analysis, protein levels were normalized to DIV0. Protein levels of PGC1α, TFAM, HSP60, Mfn1, Mfn2, OPA1, and DRP1, which are not represented as fold changes relative to DIV0, are shown in Figs S8 and S9.

Morphometric analysis

SVZ neurospheres were plated at a density of 60 neurospheres per well onto coated glass coverslip in 24-well plates. To evaluate the mitochondrial network in the different cell types, cells at DIV0, 2, 4, 7, and 14 were fixed with 4% PFA in PBS. Afterward, they were permeabilized and blocked with 0.5% Triton X-100 and 3% BSA in PBS. Cells were then incubated with primary antibodies (anti-SOX2, anti-GFAP, anti-DCX, anti-αtau, anti-NG2/anti-PDGFRα, and anti-MBP to identify NSCs, astrocytes, immature neurons, mature neurons, OPCs, and oligodendrocytes, respectively) diluted in PBS with 0.1% Triton X-100 and 0.3% BSA (wt/vol), overnight at 4°C, and then with the corresponding secondary antibodies in PBS for 2 h at RT (Table 1). Nuclei were stained with 1 μg/ml DAPI in PBS, followed by mounting with Mowiol fluorescent medium. Importantly, in addition to the differentiation process, in oligodendrocytes, the analysis was also performed accordingly to the maturation stage because of the clearly observed differences in the mitochondrial structure between less and more mature oligodendrocytes. The distinction between the two types of maturation stages was based on the complexity of the oligodendrocyte branches. Less mature oligodendrocytes presented poorly branched processes, whereas more mature oligodendrocytes presented complex branched processes. Mitochondrial network was identified through HSP60 staining. Importantly, the choice of the DIVs was based on the higher abundance of each cell type in culture. The fluorescence images were photographed on a ZEISS Cell Observer Spinning Disk confocal equipped with the ZEN software, using a 63x objective.

The neural cells were manually selected and then the morphometric analysis was performed on the mitochondria channel. For that, the Mitochondrial Network Analysis (MiNA) ImageJ macro (Valente et al, 2017) that is based on the existing ImageJ plug-in Skeleton was used to evaluate the mitochondria morphology. To obtain a sharper image with high contrast and minimal noise, the “unsharp mask,” CLAHE, and median filters were applied together with the kernel convolution. In differentiated cells, the values obtained were normalized to the cytoplasmic area. A minimum of 40 cells were acquired per condition.

Magnetic-activated cell sorting of differentiated cells

SVZ tertiary neurospheres were collected and dissociated using the NeuroCult Chemical Dissociation Kit, as previously described. Cells were then plated in coated dishes with 100 μg/ml PDL, in SFM supplemented with 5 ng/ml EGF and 2.5 ng/ml FGF-2 (low EGF/FGF-2) at a density of 1 × 10⁵ cells/cm². At 96 h of plating, differentiated cells were sorted by magnetic-activated cell sorting. For that, anti-ACSA-2 MicroBeads were used to sort astrocytes followed by incubation with anti-O4 MicroBeads to obtain oligodendrocytes. Finally, neurons were isolated from the ACSA-2^-/O4^- fraction by depletion of the magnetically labeled cells using the Neuron Isolation Kit. Then, ATP assay and respiratory assays were performed in the sorted cells 96 h + 3 d (day 7) and 96 h + 4 d (day 8), respectively. In addition, cells plated only 24 h in low EGF/FGF-2 medium were used as a control population abundant in NSCs. Moreover, cells plated for 96 h in low EGF/FGF-2 medium was also used as a control condition representing the heterogenous population presented before the magnetic separation. In both controls, cells were plated for ATP and respiratory assays. Moreover, to mimic the technical procedures to which the sorted cells were subjected, after the dissociation of tertiary neurospheres, single cells were also plated for 96 h + 3 d (day 7) in low EGF/FGF-2 medium to perform ATP assays and for 96 h + 4 d (day 8) to do respiratory assays.

Respiratory assays

Mitochondrial respiration was evaluated by measuring the OCRs performing the Mito Stress protocol. This experiment was performed in the Seahorse Extracellular Flux (XF) 24 Analyzer (Seahorse Bioscience Agilent). The 24, 96, and 96 h + 4 d CTRs were plated onto 24-well Seahorse plates at the densities of 5 × 10⁴, 1 × 10⁴, and 5 × 10⁴ cells/well, respectively. Regarding the sorted cells, astrocytes, oligodendrocytes, and neurons were seeded at the densities of 1 × 10⁵, 1 × 10⁵, and 5 × 10⁵ cells/well, respectively. The low EGF/FGF-2 medium was replaced by the XF base medium supplemented with 10 mM glucose, 2 mM L-glutamine, and 1 mM sodium pyruvate, pH 7.4. After measurements of resting respiration, cells were treated sequentially with oligomycin (2.5 μM for controls and 1.5 µM for sorted cells) to measure the nonphosphorylating OCR, two injections of FCCP (first injection: 2.5 μM for the 24-h control and 2 μM for the remaining conditions; second injection: 0.5 μM for all the conditions) to get the maximal OCR, and antimycin A and rotenone (1 μM) to measure the extramitochondrial OCR. Each measurement was taken over a 3-min interval followed by 3 min of mixing and 2 min of incubation. Three measurements were taken for the resting OCR: after oligomycin treatment, after FCCP, and after antimycin A/rotenone treatment. OCR levels were normalized to protein levels. Protein concentration of the samples was determined using the Pierce BCA Protein Assay Kit as described by the manufacturer. Importantly, the %OCR was calculated by normalizing the raw OCR values (normalized only by protein) to the first measured point, and then the obtained values were multiplied by 100 to express them as a percentage. This normalization is crucial to eliminate the inherent basal respiration variability within each replicate.

ATP content determination

To determine ATP content, a luciferase-based luminescent ATP determination assay was used according to the manufacturer’s protocol (Abcam). All ATP content levels were normalized to protein levels.

Statistical analysis

Data are represented as mean ± SEM or median ± interquartile range. Graphical illustrations and significance were obtained with GraphPad Prism 9 (GraphPad). Significance was calculated as described in each figure legend. Values of P < 0.05 were considered to represent statistical significance.