Pathogenic p62/SQSTM1 mutations impair energy metabolism through limitation of mitochondrial substrates

Abnormal mitochondrial function has been found in patients with frontotemporal dementia (FTD) and amyotrophic lateral sclerosis (ALS). Mutations in the p62 gene (also known as SQSTM1) which encodes the p62 protein have been reported in both disorders supporting the idea of an ALS/FTD continuum. In this work the role of p62 in energy metabolism was studied in fibroblasts from FTD patients carrying two independent pathogenic mutations in the p62 gene, and in a p62-knock-down (p62 KD) human dopaminergic neuroblastoma cell line (SH-SY5Y). We found that p62 deficiency is associated with inhibited complex I mitochondrial respiration due to lack of NADH for the electron transport chain. This deficiency was also associated with increased levels of NADPH reflecting a higher activation of pentose phosphate pathway as this is accompanied with higher cytosolic reduced glutathione (GSH) levels. Complex I inhibition resulted in lower mitochondrial membrane potential and higher cytosolic ROS production. Pharmacological activation of transcription factor Nrf2 increased mitochondrial NADH levels and restored mitochondrial membrane potential in p62-deficient cells. Our results suggest that the phenotype is caused by a loss-of-function effect, because similar alterations were found both in the mutant fibroblasts and the p62 KD model. These findings highlight the implication of energy metabolism in pathophysiological events associated with p62 deficiency.


Results
p62 deficiency is associated with decreased mitochondrial membrane potential (ΔΨ m ). Mitochondrial health and function are reflected in the mitochondrial membrane potential (ΔΨ m ).
Both, p62 KD SH-SY5Y cells and p62 mutant fibroblasts showed a depolarization in response to the F 0 -F 1 -ATP synthase (ATPase or complex V) inhibitor oligomycin (2 µg/ml), suggesting a reverse mode function for the ATPase ( Fig. 1E and F). Subsequent inhibition of complex I by rotenone (5 µM) caused a rapid loss of potential in all cells ( Fig. 1E and F) and a complete depolarisation was assessed by addition of the mitochondrial uncoupler carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazone (FCCP) (1 μM) ( Fig. 1E and F). These data suggest that ΔΨ m in p62 KD cells is partially maintained by ATP hydrolysis by the ATPase (Fig. 1E and F).

Discusion
Mitochondria are the main source of energy in neurons due to their limited glycolytic capacity. Harnessing the proton gradient generated in the respiration through complexes I to IV, neurons rely on the oxidative phosphorylation process to generate ATP through the ATP synthase in the mitochondria. Mutations in FTD and ALS have been found to induce impairments in crucial mitochondrial metabolic processes causing a fatal fate for cells, organs and patients 27 . Previous works presented p62 protein as one of the main regulators of mitochondrial function 14,16,17 . In this study, the mitochondrial bioenergetics in p62 deficient cells and fibroblasts from FTD patients carrying independent p62 mutations was investigated. Reduced p62 levels are associated with reduced ΔΨ m in both p62 SH-SY5Y deficient cells and mutant fibroblasts. These cells also exhibit inhibited mitochondrial respiration through complex I due to the lack of substrates for the ETC complexes and higher superoxide production. The boost in the cytosolic ROS levels leads to activation of the pentose phosphate pathway, thereby increasing the GSH levels in order to counteract the oxidative stress environment. Mitochondrial dysfunction in ALS has been thoroughly demonstrated [28][29][30] but not much is known about this topic in FTD. Recent studies found mitochondrial alterations in FTD cases together with ALS supporting the idea of the FTD/ALS continuum [18][19][20][31][32][33][34] . Other works showing mitochondrial alterations were carried out in different cohorts of patients with variations in the CHCH10 gene, a common cause of both FTD and ALS 31,35 . A number of studies have demonstrated that mitochondrial activity is compromised in FTD and ALS using animal models and cells from patients 22,36 . In this study, a reduction in the ΔΨ m is observed in the p62 deficient cell models. The ΔΨ m reflects mitochondrial health and function and mitochondrial viability and it is defined as the proton gradient generated in the intermembrane space during respiration. This proton gradient is harnessed by the ATPase to synthesise ATP through the oxidative phosphorylation (OXPHOS) process. Here it is also shown that ATP hydrolysis by the ATPase maintains the ΔΨ m in the p62 deficient cells since these cells exhibited depolarised mitochondria when oligomycin, the inhibitor of the F 1 -F 0 ATPase, was added. Under normal conditions the process of respiration maintains the ΔΨ m but mitochondria can also hydrolyze the ATP forcing the ATPase to work in a reverse mode in order to maintain the ΔΨ m 37 . This has been thoroughly described in other neurodegenerative models in which an inhibition of respiration hampers the normal ΔΨ m maintenance forcing the reversal ATPase [38][39][40] . In combination with the reduced ΔΨ m an inhibition in the respiration process was detected in p62 deficient cells. This was reflected by both the elevated NADH redox index and the decreased FAD redox index, pointing to the inhibition of complex I. This inhibition is a consequence of lack of substrates as demonstrated by the lower NADH pool observed in the p62 deficient cells as well as the lower FAD pool. The inhibition of the respiration through the complex I induces an increase in ROS production in the cells. Damaged mitochondrial respiration including reduced complex I and II activities were previously found in mitochondria from p62 KO mice brains 16 . These authors also found an exacerbated oxidative stress due to inhibition of respiration. On the other hand, in 2012 Kwon and colleagues showed mitochondrial impairments in a p62 KO model reflected in reduced mitochondrial membrane potential and reduced oxygen consumption together with increased cellular oxidant levels 17 . Increased ROS generation has been found to be related to inhibited mitochondrial respiration in neurodegenerative disorders [41][42][43] . In AD, it has been found that an inhibition in mitochondrial respiration through complex I increased ROS generation 44 . Increased ROS production is also a demonstrated feature in some forms of PD. This has been related to inhibition in the complex I dependent respiration both in cells and PD animal models 39,45,46 .
The results presented herein show that Nrf2 activators restore the inhibition of the respiration and the NADH pool in the p62 deficient cells. Nrf2 is a transcription factor implicated in the activation of antioxidant genes, providing cytoprotection against oxidative stress and inflammation 47 . More recently, new roles regarding the maintenance of mitochondrial function and bioenergetics have been attributed to Nrf2 24,47-49 . It has been shown that cells from Nrf2-deficient mice have impaired mitochondrial function. Specifically, it was demonstrated that Nrf2 is a main regulator in the substrates supply to both complex I and complex II as the rate of NADH and FADH 2 production was much lower in the Nrf2 KO cells compared to controls 24 . Nrf2 and p62 were shown to intersect through the direct interaction between p62 and Keap1 26 . Under normal conditions Nrf2 is ubiquitinated and degraded by the proteasome as it is bound to the E3-ubiquitin ligase adaptor protein Keap1. When Keap1 is challenged with ROS or electrophiles, Nrf2 degradation is blocked, activating the antioxidant response elements (ARE)-mediated transcription. In the noncanonical pathway of Nrf2 regulation, Keap1 is sequestered by p62 and can no longer bind Nrf2, leading to increased Nrf2 signalling with the subsequent activation of the ARE-mediated transcription. Our findings point to a disruption in the Keap1 sequestration by p62, leading to enhanced degradation of Nrf2. The consequence would be inhibition in respiration due to low Nrf2 levels, which could be restored by Nrf2 pharmacological activators. Finally, it is shown that p62 deficiency is associated to an increased activity in the pentose phosphate pathway as demonstrated by the elevated levels of NAD(P)H and reduced glutathione. It has been demonstrated that elevated ROS levels in cells induce the activation of the PPP to generate the antioxidant GSH 50 . The expression levels of the rate limiting enzyme responsible for GSH biosynthesis, the glutamate-cysteine ligase, are very low in neurons. This fact makes neurons more vulnerable to oxidative stress as the GSH-dependent antioxidant system is very weak in these cells. Neurons can compensate their oxidative stress vulnerability by diverting part of the glucose pool to the PPP, which generates NAD(P)H and increases the GSH levels 51 . It is possible that p62 deficient cells divert part of their glucose to the PPP to increase GSH levels after an oxidative stress stimulus, as has been shown in neurons 52 .
In summary, our results show that p62 deficiency induces mitochondrial respiration inhibition by deficient delivery of substrates to the mitochondrial complexes, which subsequently results in elevated cytosolic oxidative stress. The rise in the oxidative stress could partially trigger glucose diversion to the PPP in order to increase the GSH levels.
Even though our results were obtained using fibroblasts from FTD patients, they have been also validated using the SH-SY5Y p62 KD model, suggesting a loss-of-function effect upon p62 deficiency. This point makes the results applicable not only to FTD patients carrying p62 mutations but also to ALS patients with mutations in p62, and therefore supports the idea of FTD/ALS continuum. The data presented here along with recent works showing functional mitochondrial impairments linked to mutations causing both FTD and ALS provides significant basis to support the hypothesis that mitochondrial dysfunction is involved in the underlying pathogenic mechanisms within the FTD/ALS spectrum. Additionally, the pharmacological activation of Nrf2 needs to be further explored as therapeutic tool for the p62-associated ALS/FTD prevention and treatment.

Methods
All methods were performed following the relevant guidelines and regulations approved by the local ethical review committee from the National Hospital for Neurology and Neurosurgery and the Institute of Neurology.

Donors. Written informed consent was obtained from the donors for publication of their individual details.
The consent form is held by the authors' institution within the patients' clinical notes and is available for review by the Editor-in-Chief. Donors gave written consent and the project was approved by the local ethical review committee from the National Hospital for Neurology and Neurosurgery and the Institute of Neurology (London, UK). Fibroblasts from two patients carrying independent p62 mutations and two non-related healthy donors used as controls were generated from a 4-mm skin punch biopsy taken under local anesthetic following informed consent. Biopsies were dissected into ~1-mm pieces and cultured in 5-cm 2 petri dishes in DMEM, 10% FBS, 1% L-Glutamine until fibroblasts were seen to grow out from the explants. When fibroblasts reached confluency, they were detached from culture dishes using TrypleE from Thermo Fisher Scientific (Waltham, MA, USA) and transferred to larger culture vessels for further expansion and cryopreservation. Age, sex, age of onset, clinical features, clinical diagnosis, and p62 mutations carried by the patients as well as healthy donors characteristics are given in Table 1. Family trees from patient 1 and patient 2 are provided in (Supplementary Fig. S1).

Plasmids and reagents.
The non-targeting scramble siRNA, the targeted siRNA (siGenome SMARTpool) against human p62 and the Dharmafect transfection reagent were purchased from Dharmacon, Thermo Fisher Scientific (Waltham, MA, USA). The siRNA transfection was performed following the manufacturer's instructions once the plated cells reached 60% confluence. Cells were ready for subsequent experiments after 48 h post transfection 37 °C and 5% CO 2 .
Nrf2 activators. Three different activators for the transcription factor Nrf2 were used: the synthetic acetylenic tricyclic bis(cyanoenone) TBE-31 54 , the triterpenoid RTA-408 (RTA) 55 , and the naturally occurring isothiocyanate sulforaphane (SFN) 56 . The concentration of each compound was optimized based on the potency in inducing NAD(P)H:quinone oxidoreductase 1 (NQO1), a prototypic Nrf2 target gene, without causing any cytotoxicity, using a quantitative bioassay 57 . The compounds were prepared as stock solutions in acetonitrile and Scientific REPoRts | 7: 1666 | DOI:10.1038/s41598-017-01678-4 diluted (1:1000) in the cell culture medium, such that the final concentration of acetonitrile was maintained at 0.1% (v/v). Cells were exposed to the Nrf2 activators for 24 h. Measurement of mitochondrial membrane potential (ΔΨ m ). ΔΨ m was measured as was described previously 22 . Briefly, cells were plated on 25 mm coverslips and loaded with 40 nM tetramethyl-rhodamine methyl ester (TMRM) in a HEPES-buffered salt solution (HBSS) (composed of 156 mM NaCl, 3 mM KCl, 2 mM MgSO 4 , 1.25 mM KH 2 PO 4 , 2 mM CaCl 2 , 10 mM glucose and 10 mM HEPES; pH adjusted to 7.35 with NaOH) for 40 minutes at room temperature and keeping the dye present in the chamber at the time of recording. TMRM is a cell-permeant fluorescent dye used in the redistribution mode to assess ΔΨ m , and therefore a reduction in TMRM fluorescence represents ΔΨ m depolarization. Confocal images were obtained using a Zeiss 710 VIS CLSM (Zeiss, Oberkochen, Germany) equipped with a META detection system and a × 40 oil immersion objective. TMRM was excited using the 560 nm laser line and fluorescence was measured above 580 nm. The Z-stack images were analyzed using the Volocity software (PerkinElmer, Waltham, MA) and TMRM values for control cases were set to 100% and the p62 deficient cells values were expressed relative to controls. For analysis of response to mitochondrial toxins, images were recorded in a time course-dependent manner from a single focal plane and analyzed using ZEN Zeiss software (Zeiss).
Measurement of NADH-FAD autofluorescence. NADH autofluorescence was measured using an epifluorescence inverted microscope equipped with a 20× fluorite objective. Excitation light at a wavelength of 350 nm was provided by a Xenon arc lamp, with the beam passing through a monochromator (Cairn Research, Faversham, Kent, UK). Emitted fluorescence light was reflected through a 455 nm long-pass filter to a cooled CCD camera (Retiga, QImaging, Surrey, BC, Canada) and digitised to 12 bit resolution. Imaging data were collected and analyzed using software from Andor (Belfast, UK). FAD autofluorescence was monitored using a Zeiss 710 VIS CLSM equipped with a META detection system and a × 40 oil immersion objective. Excitation was measured using the 454 nm Argon laser line and fluorescence was measured from 505 to 550 nm. Illumination intensity was kept to a minimum (at 0.1-0.2% of laser output) to avoid phototoxicity and the pinhole set to give an optical slice of ~2 μm.
Complex I activity assay. The NADH dehydrogenase activity of isolated complex I was measured using the complex I enzyme activity microplate assay kit (Abcam, ab109721). Cell lysis was carried out using the lysis buffer provided by the manufacturer and left on ice for 20 min to allow protein extraction. Samples were then centrifuged at 12, 000 × g for 20 min at 4 °C and total protein concentration in the supernatant was estimated using the Pierce BCA Protein Assay Kit (Thermo Fisher, Waltham, MA, USA). Samples were then diluted in Incubation Solution (provided by the manufacturer) to reach a final protein concentration of 0.5 μg/μl. 200 μl of each sample were transferred to each well of the microplate containing immobilised anti-complex I antibody bound to the wells and incubated for 3 hr at RT. After the incubation period, the wells were rinsed twice in 300 μl of Buffer, provided by the manufacturer, and 200 μl of Assay Solution containing NADH and a reporter dye were added to each well. NADH dehydrogenase activity was determined by measuring the oxidation of exogenous NADH to NAD + , coupled to the 1:1 reduction of the reporter dye of which product concentration was proportional to the increase in absorbance at 450 nm, measured over time using a spectrophotometer. Complex I activity was expressed as the rate of increase in absorbance per amount of sample loaded in the well. ROS measurements. Cellular ROS generation was measured using dihydroethidium (Het, DHE; 2 μM for superoxide) from Life Technologies (Paisley, UK). All imaging was performed in HBSS and the dye was present in the solution during the experiment. No preincubation ("loading") was used for Het to limit the intracellular accumulation of oxidized products. Fluorescence images were collected with a 12-bit resolution cooled CCD camera coupled to an epifluorescence inverted microscope equipped with a 20× fluorite objective (Leica Microsystems). The excitation wavelength for the oxidised form (ethidium) was 530 nm collected at 605 nm while excitation to measure changes in the reduced form (hydroethidium) was 380 nm collecting the emission at 405-470 nm. Ratiometric Het fluorescence was recorded with excitation light at 380 and 530 nm. All imaging data were collected and analyzed using the Metamorph software (Molecular Devices, US). GSH measurements. The reduced glutathione (GSH) levels were determined in live cells after incubation with 50 μM monochlorobimane (MCB) for 40 minutes at room temperature in HBSS until a steady state was reached. MCB is a nonfluorescent probe until conjugated with GSH. Once the cytosolic GSH was analyzed, HBSS was removed and a hypotonic medium (135 mM KCl, 10 mM NaCl, 20 mM HEPES, 0.5 mM KH 2 PO 4 , 1 mM MgCl 2 , 5 mM EGTA and 1.86 mM CaCl 2 at pH 7.1) including digitonin (40μM) was added to permeabilise the cells allowing us to measure the mitochondrial GSH. The fluorescence images of the MCB-GSH adduct were acquired using the cooled CCD imaging system as described above using excitation at 380 nm and emission at >400 nm. The fluorescence was then quantified using the Metamorph analysis software (Molecular Devices, US). Additionally, quantitative GSH analysis was performed in live cells in a modified protocol as was previously described 58 . Briefly, cells were plated in 96 well plates. Once they reached 70-80% of confluence they were loaded for 30 minutes with 2 mM MCB allowing the formation of the MCB-GSH fluorescent adduct inside cells. The fluorescence was then quantified using the Varioskan plate reader (Thermo Fisher, Waltham, MA, USA) with excitation wavelength at 380 nm and emission at 470 nm. Fluorescence values were then related to the protein content (Pierce BCA Protein Assay Kit, Thermo Fisher, Waltham, MA, USA). Statistical Analysis. Data were generated from a minimum of three independent replicate per experiment (n ≥ 3) performed in different days. Each replicate consisted of at least 1 coverslip per condition where a number of 15-30 cells per coverslip were analyzed. Statistical analysis and exponential curve fitting were performed using GraphPad Prism 6.01 (GraphPad Software, La Jolla, CA) software. Statistical significance for multiple comparisons was performed by one-way ANOVA followed by Fisher's LSD correction. All results are related to healthy control fibroblasts or untransfected cells accordingly and expressed as percentage. In all cases, P < 0.05 was considered significant (*p < 0.05, **p < 0.01, ***p < 0.001). For all graphs, error bars represent mean ± SEM. Ethical Approval and Informed Consent. Primary fibroblast lines were generated from skin punch biopsies. Donors gave written consent, and the project was approved by the local ethical review committee from the National Hospital for Neurology and Neurosurgery and the Institute of Neurology.