SETD5 haploinsufficiency affects mitochondrial compartment in neural cells

Neural stem cells (NSCs) were derived from Telencephalic cortex of embryos at 14.5 days of gestation. Embryonic cortices were dissociated, fragmented in Hank’s Balanced Salt Solution (HBSS, Life Technologies) with 1% Penicillin/Streptomycin (Sigma-Aldrich) and digested with papain (10 U/ml, Worthington Biochemical) and cysteine (1 mM, Sigma-Aldrich) in HBSS with 0.5 mM EDTA at 37 °C. The obtained NSCs were routinely cultured in suspension as neurospheres.

Cells were cultured in Neural inducing medium (NIM) composed of: DMEM/F12 (Sigma-Aldrich) supplemented with Hormon Mix (DMEM/F12, 0.6% Glucose (Sigma-Aldrich) (30%), Insulin (Sigma-Aldrich) 250 µg/ml, putrescin powder (Sigma-Aldrich) 97 µg/ml, apotransferrin powder (Sigma Aldrich), sodium selenite 0.3 µM, progesterone 0.2 µM), 1 mg/ml penicillin/streptomycin (Sigma-Aldrich), 2 mM glutamine (Sigma-Aldrich), 0.66% Glucose (30% in phosphate buffer salt (PBS) (Euroclone)), Heparin 4 µg/ml, 10 ng/ml bFGF (basic fibroblast growth factor) (ThermoFisher Scientific) and 10 ng/ml EGF (epithelial growth factor) (10 ng/ml) (ThermoFisher Scientific).

To culture NSCs in adhesion, cells were seeded in coverslips coated with 1:100 Matrigel.

For low oxygen culture condition, cells were maintained in an incubator with 5% oxygen concentration. For high oxygen culture, cells were cultured in an incubator with 21% oxygen concentration.

Mutant Setd5+/− NSCs were generated as previously explained using Crispr-Cas9 technology [24].

NSCs were cultured in normal NIM using DMEM without glucose (GIBCO) instead of normal DMEM. The media was supplemented with 25 mM galactose, Hormon Mix which contains glucose was replaced by N2 supplement, the rest of the ingredients remained the same of NIM.

To perform RNA extraction, cells were seeded in adherent condition (see above) in 15 mm Petrie dish at approximately 60–70% of confluence and were grown until they reached almost 90% of confluence. Cells were then scraped and pelleted in cold (4 °C) and sterile PBS. RNA extraction was performed using RNAeasy mini kit (Qiagen).

Each sample was lysed using 600 µl of RLT buffer of RLT buffer and 1 volume of 70% ethanol was added and to the cell lysate. Then up to 700 µl of the sample, were transferred to an RNAeasy spin column placed in 2-ml collection tube and centrifuged for 15 s at 10,000 rpm and the flow trough was discarded. If the volume of the sample exceeded 700 µl the centrifuge is repeated a second time with the rest of the sample.

After this step, DNAse digestion is performed to eliminate DNA contaminants. First wash with 350 µl of RW1 is performed centrifuging at 10,000 rpm for 15 s, after the wash a mix compose of 10 µl of DNAse and 70 µl of buffer RDD are added to each column and incubation at room temperature for 15 min is performed, then after incubation another wash with 350 µl of RW1 is performed centrifuging at the same speed and time of the first wash. Buffer RPE (500 µl) is then added to the column and centrifuge for 15 s at 10,000 rpm, the flow through is then discarded. Another wash step with RPE is then performed, centrifuge step this time is for 2 min at the same speed. Then spin column is then placed on a new 2-ml collection tube and 1 min centrifuge at full speed is performed to eliminate RPE buffer carryover.

Finally, to elute the RNA a new 1.5-ml collection tube is positioned under the spin column, and 30 µl are gently posed on the filter and 1 min centrifuge is performed to elute the RNA.

The sequencing was performed by the Centre for Translational Genomics and Bioinformatics at San Raffaele Hospital. Libraries have been prepared using Tru Seq RNA v2 kit (Illumina) starting from total RNA. Sequencing was performed using HiSeq 2500 (Illumina).

Fastq files, after adaptors trimming using Trimmomatic [30], were aligned to the 10-mm mouse reference genomes using Bowtie2 [31]. Differential gene expression and Functional enrichment analyses were performed with DESeq2 [32] and GSEA [33], respectively. Statistical and downstream bioinformatics analysis were performed within the R environment. RNA-seq of NSCs cultured in low oxygen condition was performed as described in our previous work [24].

The data were deposited in the NCBI Gene Expression Omnibus and are accessible through GSE103912 (NSCs low oxygen) and GSE222188 (NSCs high oxygen).

To perform RNA extraction, NSCs were firstly seeded in adherent condition using matrigel 1:100 in 6 wells at a density of at least 200,000 per well in NIM.

After 2–3 days, medium was removed, cells were washed once with PBS and were dissolved using 1 ml of Trizol (ThermoFisher Scientific) 5 min at room temperature, directly in the well. After dissolving, samples were collected in 1.5-ml tubes (Eppendorf) and 0.2 ml of chlorophorm were added, tubes were then shaken vigorously and then incubated 2–3 min at room temperature. Tubes were subsequently centrifuged at 13,000 rpm for 15 min at 4°. An aqueous phase appears at the top of the liquid volume, that is collected and transfer to a new tube, the rest is discarded.

To precipitate RNA Isopropyl alcohol is added, 0.5 ml per 1 ml of Tryzol, shake and incubated for 10 min at room temperature before centrifuge 10 min at 13,000 rpm at 4°. After this passage, a pellet is often visible, supernatant is removed and 1 ml of 75% ethanol per ml of Tryzol is added, then samples are vortexed and centrifuged at no more than 7500 rpm for 5 min at 4°. Finally, ethanol is removed, and RNA pellet is dried at room temperature, then a suitable volume of nuclease free water is added, and samples are posed at 55° for a maximum time of 10 min to allow complete resuspension of RNA.

RNA retrotranscription was performed using ImProm II reverse transcription kit (Promega).

Briefly, the reaction is divided in two parts. The first reaction is a denaturation step of a mix composed of 1 µg of RNA, random primers (0.5 μg/reaction) and water up to 5 μl, for 5′ at 70 °C. Then after chilling the first mix on ice for 5′, a second retronscription mix (4 μl of ImProm II 5 × buffer, 3.8 μl MgCl2 (final concentration 2.5 mM), 1 μl dNTP Mix (final concentration 0.5 mM each dNTP), 1 μl ImProm II reverse transcriptase, 5.2 of nuclease free water) is added to first mix, and then reverse transcription is performed with the following reaction profile:

After c-DNA synthesis real-time quantitative PCR was performed (RT-qPCR) using CFX 96 Real-Time System (Biorad) thermocycler.

The real-time TITAN (Bioatlas) mix was composed as follows for each reaction samples:

The amplification profile used was the following:

RT-qPCR data were analyzed using CFX Manager software (Biorad) and the differential gene expression was calculated using the ∆∆ct methods. Used primers are listed in the Additional file 3: Table S3.

NSCs were plated on Matrigel-coated glass coverslips (13 mm) and were fixed for 20 min on ice in 4% paraformaldehyde (PFA, Sigma), solution in phosphate-buffered saline (PBS, Euroclone). Two washes were performed afterwards with PBS, cells were then permeabilized for 30 min in blocking solution, containing 0.1% Triton X-100 (Sigma-Aldrich) and 5% donkey serum (Sigma-Aldrich), and incubated overnight at 4 °C with antibody against transporter of outer mitochondrial membrane 20 (TOMM-20) diluted in blocking solution. The next day, cells were washed 3 times with PBS for 5 min and incubated for 1 h at room temperature with Hoechst 33,342 (ThermoFischer Scientific) and specific secondary antibodies (ThermoFisher Scientific) in blocking solution.

Images were acquired using confocal microscope (Leica Sp5). Fragmentation level was assessed based on circularity calculated using ImageJ plugin mitochondrial morphology (http://​imagejdocu.​tudor.​lu/​doku.​php?​id=​plugin:​morphology:​mitochondrial_​morphology_​macro_​plug-in:​start).

For adult mice, the brain of wild type and Setd5+/− were extracted from the skull after brief perfusion with Sodium Chloride 0.9% (S.A.L.F) and then included in TEM fixative solution for further analysis.

Major axis of mitochondria was measured using ImageJ software. Qualitative evaluation of mitochondria was performed manually on each image.

To perform Western Blot analysis cells were seeded in adherent condition in 15 mm Petrie dish till 90% of confluence. Cells were lysed in RIPA buffer (50 mM Tris–Hcl ph 7.4, 150 mM NaCl, 0.1% SDS, 1% Triton X-100, 2% complete protease inhibitor cocktail (Roche), 10%complete phosphatase inhibitor cocktail) for 30 min at 4°. After the lysis samples were centrifuged at 13,000 rpm for 30 min and the supernatant was rescued. Protein samples were then quantified using BCA protein assay kit (ThermoFisher).

After quantification protein samples containing 30 µg of protein, sample buffer (Tris–Hcl ph 6.8 150 mM, glycerol 45%, SDS 7,50%, Bromophenol blue 0.08%) and H₂O up to 24 µl each were prepared and boiled for 5 min at 95°.

Samples were loaded on poly-acrylamide gel and after proper run were transferred on nitrocellulose membranes for 90–120 min at constant intensity of 350 mA.

Blocking reaction was performed using 5% milk in PBS-T (PBS with 0.1% tween) for 1 h at room temperature. Incubation with primary antibody (Additional file 3: Table S3) was performed over/night at 4° in blocking solution at the proper dilution. The next day membranes were washed 3 times, for 15–20 min in PBS-T and incubation with secondary antibody (Additional file 3: Table S3) conjugated with horseradish peroxidase (HRP) was performed at the proper dilution in blocking solution for 1 h at room temperature.

Finally, to reveal the signal ECL prime kit (GE healthcare) has been used and images were acquired using a chemiluminescent reader.

Densitometry was calculated using ImageLab software (Biorad).

NSCs were analyzed by fluorescence activating cell sorter (FACS) using LSRFortessa analyzer (BD Bioscience), using two different fluorescent dyes, MitoTracker^® Green FM (ThermoFisher) and MitoTracker^® Orange CMTMRos (ThermoFisher). The first one stains mitochondria regardless of the membrane potential, the second one stains mitochondria depending on the membrane potential (Fig. 2). Using both at the same time allows to determine if there is a difference in the membrane potential and in the total mitochondria mass in the whole cell population.

Cells were seeded at a density of approximately 500,000 cells per well in 6 wells in adherent condition the day before the analysis. The next day cells were stained with MitoTracker Green FM (stock 1 mM in DMSO) and MitoTracker Orange CMTMRos (stock 1 mM in DMSO) at a final concentration of 50 nM diluted in NIM in both cases. Each sample was single and double stained with both dyes.

To perform live imaging on NSCs with TMRM, cells were seeded in adherent condition on glass coverslips at a density of 30,000 cells per well in 24 wells plate. The die was diluted at a final concentration of 250 nM and then cells were incubated at 37° for 10′ or 20′ before image acquisition.

To measure the maximum level of membrane potential cells were treated with complex V inhibitor oligomycin at concentration of 1 µM. To measure the background level of TMRM fluorescence, NSCs were treated with decoupling agent FCCP at concentration of 0.7 µM.

After incubation with the fluorescent dyes, coverslips were mounted on cover glasses using mounting medium, without performing fixation and images were immediately acquired using epifluorescence microscopy. Images were quantified and analyzed using ImageJ software, using gray scale to quantify the fluorescence intensity of the cells for TMRM.

Cellular ATP concentration was determined using ATP-Lite assay kit (Perkin-Elmer), an assay that use luminescence to determine cellular ATP concentration.

Cells were seeded in 96 wells plate coated with Matrigel 24 h before the experiment at a density of 30,000 cells per well. The next day cells were lysed using the Mammalian cells lysis solution provided by the kit, then the plate is shaken at 700 rpm for 5 min. The lysates are then transferred to a dark adapt plate and the substrate solution is added and the plate is shaken for 5 min avoiding light exposure, then after ten minutes the plate can be read using a luminometer.

Lactate concentration in the medium were measured EnzyChrom™ L-Lactate Assay Kit. Cells were seeded in a Matrigel-coated 96 wells plate at a density of 30,000 cells per well. After 72 h 20 µl of medium are taken for each well and 80 µl (60 µl of assay buffer, 1 µl of Enzyme A, 1 µl of Enzyme B, 10 µl of NAD and 14 µl of MTT). Optical density (OD) is read at 565 nm at time 0 and at time 20, 20 min after adding the working solution. OD values obtained at time 0 are subtracted to time OD obtained at time 20 and it’s possible to calculate the lactate concentration building a standard curve using, the range of detection varies depending on the medium or samples that is measure. For cell cultured in phenol red containing media, the range of concentrations detectable spans from 0.1 to 1 mM.

To quantify the total intracellular concentration of NAD/NADH, NSCs were seeded in a Matrigel-coated 96 wells plate at a density of 30,000 cells per well 24 h before the experiment.

The NAD/NADH-Glo™ Assay (Promega) kit was used according to the manufacturer’s instructions.

To perform live imaging on NSCs with H2DCFDA and DHE, cells were seeded in adherent condition on glass coverslips at a density of 30,000 cells per well in 24 wells plate. Both dyes were diluted in PBS at a concentration of 10 µM for H2DCFDA and 3.2 µM for DHE, together with Hoechst to stain nuclei. NSCs were then incubated 15′ for H2DCFDA and for 10′ for DHE at 37°. After incubation with the fluorescent dyes, coverslips were mounted on cover glasses using mounting medium, without performing fixation and images were immediately acquired using epifluorescence microscopy. Images were quantified and analyzed using ImageJ software, using gray scale to quantify the fluorescence intensity of the cells for H2DCFDA, or the nuclei for DHE.

Cortical Neurons were obtained by dissecting telencephalic cortex of embryos, wild type and Setd5+/−, at day 17.5 post-conception from wild type and Setd5+/−. Briefly, after dissection, the cortex was digested 30′ with trypsin. After 3 washes with HBSS medium, the tissue was gently dissociated mechanically by using pipette, DNAse was added to avoid cell clumping and cells were centrifuged 5′ at 1200 rpm. Subsequently, cells were plated at density of 400,000 in an Ibidi, µ-Dish 35 mm, high chamber for live imaging coated the day before with poly-L lysine. The next day cells were infected with CMV-mito-dsRed lentivirus. After 7 days in culture, live imaging was performed using the in vivo imaging setup of Leica sp8 confocal microscope.

For Mdvi-1 rescue experiment, the compound was resuspended in DMSO as vehicle and added directly to the culture media at final concentration of 25 um the day after plating. The treatment was maintained throughout the whole culture period before performing image acquisition.

Mitochondrial fragmentation level was assessed based on circularity calculated using ImageJ plugin mitochondrial morphology (http://​imagejdocu.​tudor.​lu/​doku.​php?​id=​plugin:​morphology:​mitochondrial_​morphology_​macro_​plug-in:​start). Other measures were performed using image J software.

Mitochondria were isolated as previously published [34, 35], from the cortex of 3 month old wild-type and Setd5+/− mice. In brief, mice cortex was homogenized in an appropriate isotonic buffer [0.25 m sucrose, 20 mm 3-(N-morpholino) propane sulfonic acid (MOPS), pH 7.2, 1 mm EDTA, 0.1% BSA fatty acid free and digitonin 0.1 mg/ml] using a glass-Teflon homogenizer. Cell debris and nuclei were pelleted twice by centrifugation at 2500 g for 10 min at 4 °C. Supernatants were centrifuged at 12,500 g for 30 min at 4 °C and the mitochondrial pellet was resuspended in an isotonic buffer (0.5 m sucrose, 20 mm MOPS, pH 7.2, 1 mm EDTA). Isolated mitochondria were incubated at 37 °C for 30 min in a respiratory buffer (0.25 m sucrose, 20 mm MOPS, 1 mm EDTA, 5 mm inorganic phosphate, 0.1% BSA fatty acid free, and 1 mm ADP, pH 7.4) containing specific substrates of the respiratory chain complexes. By providing pyruvate/malate (5 and 1 mm, respectively) and glutamate/malate (5 and 1 mm, respectively), we stimulated ATP synthesis. ATP production was measured by luminometric assay.

Synaptosomes were isolated from cortical brain tissue of Setd5^+/+ and Setd5^+/− by using the Syn-PER reagent (Thermo Fisher). Briefly, the cortical tissue was isolated and disaggregated into a glass Dounce homogenizer filled up with 4 mL of Syn-PER reagent. After 15 gently strokes, the homogenate was poured in a 15 mL falcon and centrifuged at 1200× g for 10 m at 4 °C. The supernatant was then centrifuged at 15,000× g for 20 m at 4 °C. The pellet containing the purified synaptosome fraction was weighted for subsequent treatments.

In experiments with more than two TI, one-way analysis of variance (ANOVA) and Dunn multiple comparison test was performed to determine the statistical significance of the experiments.

In experiments with more than one factor analyze and more than two TI, two-way analysis of variance (ANOVA) and Sidak, Bonferroni or Tuckey multiple comparison test was performed to determine the statistical significance of the experiments.

Murine Setd5 haploinsufficient neural stem cells (NSCs) normally cultured at 5% O₂, to respect the condition found in the embryo and stem cells niches, showed wide downregulation of mitochondrial-associated genes as well as the increase of response of hypoxia including glycolysis by RNA-seq (Additional file 1: Table S1, Additional file 4: Fig. S1A, B) [24]. We reasoned that culturing the NSCs upon ambient air condition (21% O₂) possibly uncover mitochondrial impairment in Setd5 haploinsufficiency (Additional file 4: Fig. S1C). We found that high oxygen further exacerbated the deregulation of the mitochondrial gene subset (Fig. 1A, Additional file 1: Table S1, Additional file 4: Fig. S1D, E) being the mutant cells unable to properly react to the oxygen increase, at least at the transcriptional level (Fig. 1B). Interestingly, ~ 30% of the mitochondrial-associated genes were among the SETD5 target genes identified by ChIP-seq in the same cells (Fig. 1C, Additional file 2: Table S2) [24]. Of note, the SETD5 binding on gene bodies was particularly relevant on mitochondrial genes compared to both all genes and a random subset of the same size as indicated by the density plot (Fig. 1D, Additional file 2: Table S2). In accordance with the showed SETD5 binding preference, the mitochondrial subset encloses genes that are both rich in exons, highly expressed and long, features that emerged as correlated with a high probability of SETD5 genomic association (Fig. 1E) [24].

Together these observations highlight the importance of correct levels of SETD5 to promote the transcription of mitochondrial-associated genes especially when those genes are specifically required, i.e., in response to high O₂ tension.

Since the observed transcriptional deregulation, we investigated the mitochondrial compartment in both wild-type (WT) and Setd5 mutant NSCs upon high O₂ levels. Initially, we evaluated the fusion/fission dynamic through both immunocytochemistry and electron microscopy. TOMM20 immunostaining indicated that mitochondria in Setd5 mutant NSCs were more fragmented compared to the WT (Fig. 2A) while transmission electron microscopy confirmed the smaller size of mitochondria and showed the altered ultrastructural morphology with a high number of organelles with aberrant cristae structure (Fig. 2B). Interestingly, among genes downregulated in our RNA-seq dataset, we found genes encoding for important regulators of mitochondrial dynamics (Mfn1 and Opa1) and biogenesis (Pgc1α, Polg) that indeed were decreased in mutant cells compared to the WT (Additional files 4, 5: Figs. S1C, E, S2A).

Then, we evaluated the amount of several mitochondrial proteins as a proxy of the complexes of oxidative phosphorylation (OXPHOS), also known as the electron transport chain, which is the biochemical core of the mitochondrial functionality. In line with our transcriptional (Fig. 1) and structural data (Fig. 2A, B), Setd5 mutant mitochondria were deficient in many protein subunits of the different complexes (Fig. 2C). Since the OXPHOS complexes I, II, and IV are proton pumps, we sought defects in the generation of mitochondrial membrane potential (ΔΨm) in mutant cells. Using cell-permeant fluorescent dyes that are either insensitive (Mitotracker green) or sensitive (Mitotracker orange and TMRM) to mitochondrial membrane potential we showed, as expected, lower ΔΨm in Setd5 mutant cells compared to control despite the total mitochondrial masses were equal (Fig. 2D, E, Additional file 5: Fig. S2B). Mitochondrial membrane potential (measured as Tetramethylrhodamine, Methyl Ester, Perchlorate (TMRM) signal) was then challenged by adding to the medium chemical compounds to induce either hyper- (using oligomycin) or de-polarization (carbonylcyanide-p-trifluoromethoxyphenylhydrazone (FCCP)) [36]. Interestingly, the mutant cells were not able to properly increase their potential, having only a mild response to oligomycin while the uncoupling between the respiratory chain and phosphorylation system due to FCCP was effective (Fig. 3A, Additional file 6: Fig. S3A). Low ΔΨm and blunt reaction to the hyperpolarizing agent would suggest an impairment to generate ATP by OXPHOS. We measured a decrease in total ATP content in mutant cells compared to the control, supporting an energetic defect due to Setd5 haploinsufficiency (Fig. 3B). Accordingly, the levels of nicotinamide adenine dinucleotide (NAD) were decreased (Fig. 3C). However, the mitochondria of mutated cells seemed not to experience electron leakiness since the evaluation of reactive oxygen species (ROS) resulted in diminished in the mutant cells (Additional file 6: Fig. S3B, C).

We hypothesized that low SETD5 levels may hamper the mitochondrial maturation, and thus the related energy supply, forcing the cells to use glycolysis to survive, as the transcriptional profile already suggested (Fig. 1, Additional file 4: Fig. S1). To investigate this issue, we measured the lactate, a byproduct of glycolysis, that was indeed increased in mutant NSCs (Fig. 3D). To further analyze this aspect, we have grown both WT and mutant NSCs in a culture medium in which the glucose was replaced by galactose as the carbon source for metabolic reaction; the galactose can enter in the glycolytic pathway but without a net gain of energy; moreover, it cannot enter in the pyruvate-lactate anaerobic pathway. Thus, forcing the cells to rely only on mitochondrial respiration, to produce ATP [37]. WT NSCs behaved similarly in glucose and galactose media while the mutant cells, that with glucose proliferated more than the WT, tended to die (Fig. 3E).

These data indicate that low SETD5 levels lead to metabolic reprogramming due to an inefficient mitochondrial dynamic, maturation, and function.

To better circumstantiate our findings in regard to neurological pathologies associated with SETD5 mutations, we moved to analyze Setd5^+/− animal model. We first evaluated the abundance of OXPHOS complexes. In line with NSCs data, we showed that Setd5^+/− brains presented a decrease in OXPHOS proteins particularly evident in complex I (Additional file 7: Fig. S4A). Evaluation of ATP content in freshly isolated mitochondria from control and Setd5^+/− brains sustained the in vivo hypo-functionality of oxidative phosphorylation since isolated mitochondria from mutants produced less ATP than their normal counterpart, specifically using glutamate as the substrate for mitochondrial respiration (Fig. 4A).

Using electron microscopy, we retrieved small mitochondria in mutant brains compared to control littermates possibly confirming the defects in fission/fusion dynamics seen in NSCs (Fig. 4B). To better analyze this aspect, we moved to cultured primary neurons. Live staining using the mito-dsRed mitochondrial marker, showed that neurons from E17.5 Setd5^+/− cortices displayed more fragmented mitochondria compared to controls (Fig. 4C). Interestingly, in mutant neurons, the mitochondria were located primarily in cell soma while were reduced in neurites (Fig. 4C). We reasoned that mitochondria may be stacked in the cell body, with few possibilities to travel along neurites to reach synaptic sites within neurons. To investigate this issue, we isolated synaptosomes containing fraction from Setd5^+/+ and Setd5^+/− cortices and estimated mitochondria content showing that mutant synapses contained fewer mitochondria (Fig. 4D).

The quinazolinone derivative mitochondrial division inhibitor 1 (Mdiv1), reversibly inhibits DRP1, a master regulator of fission machinery [38]. Its action has been linked with increased mitochondrial fusion, attenuation of apoptosis, and cytoprotection [38‐42]. We decided to use Mdiv1 in treating Setd5 mutant neurons. Mdiv1 was effective to increase mitochondrial length in mutant neurons as well as mitochondrial distribution inside the mutant cell soma and neurites (Fig. 5A).